bz media publication - sd times · support popular file types, emails ... † tips and tricks for...

A BZ Media Publication

JANUARY 2017 • ISSUE NO. 333 • $9.95 • www.sdtimes.com

SDT333 cover_Layout 1 12/22/16 2:26 PM

http://www.sdtimes.com

Indiawww.MelissaData.in

Australiawww.MelissaData.com.au

United Kingdomwww.MelissaData.co.uk

Germanywww.MelissaData.de

www.MelissaData.com 1-800-MELISSA

Data Quality Made Easy.Your Data, Your Way.

Start Your Free Trialwww.MelissaData.com/sd-times

@

NAME

Our data quality solutions are available

on-premises and in the Cloud – fast, easy

to use, and powerful developer tools,

integrations and plugins for the Microsoft

and Oracle Product Ecosystems.

Melissa Data provides the full spectrum of data

quality to ensure you have data you can trust.

We profile, standardize, verify, match and

enrich global People Data – name, address,

email & phone.

SDT333 Full Page Ads_Layout 1 12/21/16 9:15 AM

http://www.melissadata.com/sd-times

NEWS 6 News Watch

18 2017 trends in software development

20 Top retailers are open to hacks

23 Android 7.1: What you can expect

24 HashiCorp and the state of automation

26 Notes from Node.js Interactive

29 The Software Testing World Cup chronicles

33 CollabNet enters DevOps arena

39 Red Hat powers the API economy

COLUMNS 48 GUEST VIEW by Alan Ho

Diffusing the monolith time bomb

49 ANALYST VIEW by Al Hilwa

Immigration, progress and politics

50 INDUSTRY WATCH by David Rubinstein

2017: The future starts now

Contents ISSUE 333 • JANUARY 2017

2016: The Year of Artificial Intelligence

Digital transformation is essential to thefuture of business

page 34

Move fast while avoiding automated testing pitfalls

page 41

page 10

Software Development Times (ISSN 1528-1965) is published 12 times per year by BZ Media LLC, 225 Broadhollow Road, Suite 211, Melville, NY 11747. Periodicals postage paid at Hunting ton Station, NY, andadditional offices. SD Times is a registered trademark of BZ Media LLC. All contents © 2017 BZ Media LLC. All rights reserved. The price of a one-year subscription is US$179 for subscribers in the U.S., $189 inCanada, $229 elsewhere. POSTMASTER: Send address changes to SD Times, 225 Broadhollow Road, Suite 211, Melville, NY 11747. SD Times subscriber services may be reached at [email protected].

FEATURES

SDT333 page 3_Layout 1 12/21/16 1:45 PM

Instantly Search Terabytes of Text

®

dtSearch.com 1-800-IT-FINDS

dtSearch’s document filters support popular file types, emails with multilevel attachments, databases, web data

(See site for articles on faceted search, SQL, MS Azure, etc.)

Highlights hits in all data types; 25+ search options

The Smart Choice for Text Retrieval®

since 1991

Visit dtSearch.com for

case studies

developer evaluations

ADVERTISING SALES

EDITORIAL

www.sdtimes.com

EDITOR-IN-CHIEF David Rubinstein

631-421-4158 x105 • [email protected]

SENIOR EDITOR Alex Handy • [email protected]

SOCIAL MEDIA AND Christina Cardoza • [email protected] ONLINE EDITORS Madison Moore • [email protected]

COPY EDITOR Adam LoBelia • [email protected]

SENIOR ART DIRECTOR Mara Leonardi • [email protected]

CONTRIBUTING WRITERS Patrick Hynds, Lisa Morgan,

Alexandra Weber Morales

CONTRIBUTING ANALYSTS Rob Enderle, Michael Facemire, Mike Gilpin,

Mike Gualtieri, Jeffrey Hammond, Al Hilwa

SUBSCRIPTIONS [email protected]

ADVERTISING TRAFFIC Mara Leonardi

[email protected]

LIST SERVICES Shauna Koehler

[email protected]

REPRINTS Stacy Burris


ACCOUNTING Viena Ludewig


OTHER QUESTIONS [email protected]

CUSTOMER SERVICE

PRESIDENT Ted Bahr

FOUNDING EDITOR Alan Zeichick

BZ MEDIA LLC

225 Broadhollow Road, Suite 211

Melville, NY 11747

TEL 631-421-4158

FAX 631-421-4130

www.bzmedia.com

[email protected]

PUBLISHER David Lyman

978-465-2351 • [email protected]

WESTERN U.S., WESTERN CANADA, EASTERN ASIA, AUSTRALIA, INDIA

Paula F. Miller


EASTERN U.S., EASTERN CANADA, EUROPE, MIDDLE EAST, WESTERN ASIA

Jonathan Sawyer


PUBLISHING DIRECTOR Ted Bahr



http://www.dtsearch.com

http://www.theleonardifamily.com

www.sptechcon.com

April 2-5, 2017 • AUSTIN, TEXAS

The BestSharePoint 2016 andOffice 365 Training!

SPTechCon offers classes and tutorials for IT professionals, business decision makers, information workers, developers and software and information architects. Each presenter at SPTechCon is a true SharePoint expert, with many drawn from Microsoft’s tech teams or holding Microsoft MVP status.

Whether you’re looking to upgrade to a more current version, making a move to the cloud, or simply need answers to those daunting problems you’ve been unable to overcome, SPTechCon is the place for you! Come join us!

• Practical information you can put to use on the job right away!

• Choose from more than 80 classes and panel sessions

• Improve your skills and broaden your knowledge of Microsoft's collaboration and productivity software

• Learn about SharePoint 2016, the latest on-premises server release from Microsoft

• Tips and tricks for working with SharePoint2013 and 2010, and Office 365

• The most knowledgeable instructors working in SharePoint today

A BZ Media Event

The LargestIndependentSharePointConference!


http://www.sptechcon.com

Yelp offers up Kafkatools to open sourceYelp saved itself US$10 million

by building out its Apache Kaf-

ka-based Data Pipeline, and

now it wants to spread that

love to other enterprises. Just

before the holidays, Yelp

open-sourced its Data Pipeline

and assorted utilities used to

maintain and build out this

streaming data platform.

Data Pipeline is now avail-

able on GitHub under the

Apache 2.0 license. Using

Data Pipeline, developers can

tie their applications into the

constantly flowing stream of

Kafka data.

Jason Fennell, vice presi-

dent of engineering at Yelp,

said that Data Pipeline cou-

pled with Kafka provides ben-

efits to all data streaming

through the company’s sys-

tems. “We’ll build a connector

from Kafka to our Salesforce

instance, and now we have

this real-time stream of

updates from our core data-

bases into Salesforce. We

managed to make a process

that could take as long as

three weeks to get data down

to a few seconds,” he said.

GrapeCity enters theJavaScript marketwith SpreadJSGrapeCity is giving JavaScript

developers a new way to pres-

ent and interact with their

data in the latest release of its

enterprise spreadsheet tool.

Spread Studio 10 comes with

SpreadJS, a JavaScript spread -

sheet and data presentation

solution.

Spread Studio version 10

also features new charts, for-

mula extenders, touch toolbar,

and support for Microsoft

Edge and Google Chrome.

The company also recently

released ActiveReports 11, its

.NET reporting solution. The

update targets both ActiveRe-

ports Developer and ActiveRe-

ports Server components. The

latest features include perform-

ance improvements, a faster

report rendering engine, native

support for JSON and CSV data

sources, a query builder for the

XML data provider, and a new

Excel Import tool.

CA Technologiesacquires AutomicCA Technologies continues to

add to its cloud-based capabil-

ities with its recent acquisition

of Automic, a European busi-

ness automation software

company. The transaction,

according to CA Technologies,

is valued at approximately

600 million euros, and it has

been approved by both boards

of directors. The transaction is

expected to close in the fourth

quarter of 2017.

This acquisition comes on

the heels of CA’s acquisition of

BlazeMeter, the continuous

application performance test-

ing company, where it inte-

grated with CA’s Continuous

Delivery solutions to offer cus-

tomers a diverse set of testing

capabilities.

With the acquisition of

Automic, CA will add the com-

pany's automation and orches-

tration capabilities to its portfo-

lio so customers can have more

options that will address their

IT and DevOps challenges,

especially for on-premises, the

cloud, and hybrid cloud envi-

ronments, said Ayman Sayed,

president and chief product

officer at CA Technologies.

OpenAI opens new AI universeOpenAI is giving artificial intel-

ligence researchers a new way

to test and evaluate their

research. The organization has

announced Universe, a soft-

ware platform designed to train

and measure the general intel-

ligence of AI across games,

websites and applications.

Researchers will also be able

to turn their programs into Gym

environments. (Gym is OpenAI’s

recently launched toolkit for

reinforcement learning.)

Environments currently

include Atari games, Flash

games, extracting rewards,

browser tasks, browser inter-

actions, and real-world brows-

er tasks.

OpenAI plans to release a

“transfer learning bench-

mark” to help researchers

determine the progress on

their experiments.

OpenMake open-sources ARA solutionOpenMake Software wants to

improve how developers use

the Continuous Delivery

pipeline with its recently open-

sourced Application Release

Automation (ARA) solution,

Release Engineer, which is

based on version 7.7 of the

ARA solution and offered

under the FreeBSD license.

According to cofounder of

OpenMake Tracy Ragan, the

ARA solution is fully function-

al and will have an upgrade

path for things like security

roles, user groups, folders,

audit tracking, and release

training management, which

are all strategic features that

are specific to what an enter-

prise would need.

Those features are avail-

SD Times January 2017 www.sdtimes.com6

NEWS WATCHNEWS WATCH

Google open-sources DeepMind LabGoogle is open-sourcing its flagship AI 3D platform for agent-based AI research, DeepMind Lab.

According to Google, the research done at DeepMindhas been to develop intelligent agents with advanced cog-nitive skills, as well as providing an environment wherethose agents can be trained and analyzed. Those simulatedenvironments are meant to be AI labs for research.Researchers can also take advantage of the lab’s program-matic level-creation interfaces. The customizable levelscan include gameplay logic, observations and the ability topick up items.

DeepMind Lab, along with its code, maps and level scripts,will be available on GitHub.

continued on page 8 >

SDT333 page 6,8_Layout 1 12/20/16 11:59 AM

DOMAINS | MAIL | HOSTING | eCOMMERCE | SERVERS

STOP SHARING!

®

1and1.com

CALLSPEAK WITH AN EXPERT 24/71 MONTH

FLEXIBLE PAYMENT OPTIONS1 SECURE

FAIL-SAFE1CHOICE

* 1&1 Virtual Server Cloud S: $4.99/month. Billing cycle 1 month. Minimum contract term 12 months. No set up free. © 1&1 Internet Inc. 2016 All rights reserved.1&1 and the 1&1 logo are trademarks of 1&1 Internet SE, all other trademarks are the property of their respective owners. Windows and the Windows logo are registered trademarks of Microsoft Corporation in the United States and/or other countries. 1&1 Internet Inc, 701 Lee Road, Chesterbrook, PA 19087.

1-844-296-2059

NEW1&1 eliminates the "noisy neighbor effect": Ideal for beginners as a web and mail server,

but also for more demanding projects like database applications, the new 1&1 Virtual Server

Cloud is 100% yours to use! Take advantage now of the latest cloud technology.

No shared resourcesthrough VMware virtualization Full root access SSD storage Unlimited traffi c High performance

Maximum security Best price-performance ratio 24/7 expert support Choice between Linux/Windows Plesk ONYX

1&1 VIRTUAL SERVER CLOUD

starting at $4.99per month*

Trusted Performance.

Intel® Xeon® processors.

SDT333 Full Page Ads_Layout 1 12/20/16 12:17 PM

http://www.1and1.com

NEWS WATCHNEWS WATCH


able in the Enterprise Upgradeto the open-source edition,while the base version will begood for development teams,said Ragan. With that version,which is hosted via GitHub,developers can build out asolution that can be con-sumed by the production envi-ronment and then tested inreal production, she said.

ARA aims to strengthen theContinuous Delivery pipelineby allowing deploymentsthrough it, from developmentthrough production. Ragansaid developers tend to writemore lightweight agileprocesses, and other offeringstypically are heavy and requirean end-target agent on everymicroservice, Docker contain-er, physical server, etc. Whathappens in the ContinuousDelivery pipeline is developersdo the turnover on the opera-tion side, where the team isforced to have a discussionabout how they are going to beable to do the deployments,according to her.

JNBridge releasesJNBridgePro 8.0JNBridge is giving developers anew way to work with softwareframeworks in JNBridgePro8.0. JNBridgePro is the compa-ny’s general-purpose Java and.NET interoperability tool.

According to Wayne Citrin,CTO of JNBridge, this fills ahuge gap and addresses a prob-lem that developers have beenfacing. While developers can’tmodify software frameworkcode, they can extend itthrough additional code andoverriding behaviors. However,if a developer wanted to previ-ously write .NET code to extenda Java framework, there arelimitations and involve time-consuming workarounds to

change underlying code or add

new components. With this

release, JNBridge supports a

natural way of doing develop-

ers, he explained.

In addition, the company

made changes to JNBridge-

Pro’s user interface providing a

flatter and more modern look.

ActiveStateannounces ActiveGoActiveState has announced

that it would be solving one of

the biggest problems enter-

prises still have with Go.

ActiveGo will be available

in beta form in February, said

Jeff Rouse, director of prod-

uct management at Active-

State, and the package will

offer long-term support for

the Go environment. That

means developers in enter-

prises will be able to back up

their Go applications with cor-

porate support and services.

Rouse said that enterprises

have trouble moving at the

speed of a new language like

Go. The biggest problem can

simply be keeping up with

fast-moving libraries. To that

end, ActiveGo includes pack-

ages that will be supported

long term in their current ver-

sions, even if the upstream

packages advance quickly in

the coming years.

Additionally, the ActiveGo

package includes crypto-

graphic packages and data-

base connectivity helpers. The

Komodo IDE is also available

to use with the Go language,

with ActiveState planning to

enhance this support even

further over time.

Google evaluatesFIDO authenticationFor years the FIDO Alliance

has been dedicated to chang-

ing and improving online

authentication. FIDO, which

stands for Fast Identity Online,

envisions a future where

online security methods go

beyond passwords and pro-

vide stronger authentication

solutions such as biometrics

and second-factor solutions.

Google recently did a two-

year research study on FIDO’s

approach to examine how well

it worked.

FIDO-based Security Keys

are devices designed to make

two-step verification more

secure and easier to use. “Our

system design goals required

Security Keys to be easy to use;

easy for developers to integrate

with a website via simple APIs;

non-trackability to ensure pri-

vacy; and protect users from

password reuse, phishing, and

man-in-the-middle attacks,”

said the company. “The cur-

rently most common version of

our Security Key is a tiny don-

gle that plugs into a computer’s

USB port, although the Security

Key’s underlying protocols are

standardized and can also be

used via NFC (contactless) and

Bluetooth Low Energy.”

The company compared

the Security Keys against one-

time password generators and

two-step SMS verifications

looking at usability, deploya-

bility, and security. According

to the company’s results, FIDO

Security Keys proved to be the

most secure as well as the

easiest to use and deploy. z

Report: Wearables need to be more rewardingWearables are giving businesses new opportunities for innovation, but in order to be truly suc-cessful, they have to go beyond the hype and provide users more capabilities. A recent reportrevealed interest in wearables is waning because they aren’t valuable enough to users.

In Gartner’s user survey analysis, the organization found the abandonment rate for smart-watches is 29%, while fitness trackers are at 30%, because users aren’t finding them useful.

The survey looked at responses from 9,592 online participants from Australia, the U.S. andthe U.K to gauge their perception of wearables and their pur-chasing habits. The report also revealed smartwatches are stillin their infancy, while fitness trackers have reached early main-stream usage. According to the report, 34% of fitness trackersand 26% of smartwatches are given as gifts.

In addition, the survey revealed people think wearabledevices are overpriced; a majority of people who own fitnesstrackers wear them every day; users under 45 years old believea smartphone is all they need; users over 45 years old don’t planon purchasing a fitness tracker because of the expense; andpeople 44 years old and younger use smartwatches the most.Gartner believes unknown wearables providers will have a hard-er time against popular brands.

< continued from page 6

SDT333 page 6,8_Layout 1 12/20/16 12:00 PM

To help bring your software vision to life, Epson is with you all the way. With our industry-leading development tools

and technical support, you get all the backup you need from the leader in POS printing. From traditional POS to

mobile to cloud, we’ve got you covered with world-class APIs and SDKs. We’ll even do customization work, and we’re

here to quickly resolve any development and technical issues you have. Whatever products and solutions you need,

we’ve got ‘em. Rock on. epson.com/mPOS

EPSON is a registered trademark and EPSON Exceed Your Vision is a registered logomark of Seiko Epson Corporation. Copyright 2016 Epson America, Inc.

Every rock star needs a great band.Epson is right behind you with world-class development and technical support.

Another Innovation from Epson Business Solutions

SDT333 Full Page Ads_Layout 1 12/21/16 1:04 PM

http://www.epson.com/mPOS


2016: The Year of Artificial Intelligence

BY DAVID RUBINSTEIN

Changes in software development and deliveryare driving a digital transformation, and AI will drive that transformation

SDT333 page 10-11_Layout 1 12/20/16 3:53 PM

www.sdtimes.com January 2017 SD Times 11

In the software development industry, 2016 was truly transformative. Theway software is developed, what it is being created for, and where itresides and is used all changed to a large degree last year. More develop-

ment teams adopted agile and DevOps techniques, while QA “moved left”and integrated into the process, rather than being a post-development effort.

The definition of the application life cycle changed (or expanded, depend-ing on your point of view) to include deployment. Development shifted fromdoing nightly builds to Continuous Integration. Software release went froma major one-time event to Continuous Delivery. Testing struggled to staywith the pack, with the realization that automated testing would be the onlyway to keep up with the quickening pace of development.

And we saw more software being deployed as a service in 2016, living oncloud servers with functionality that’s just an API call away for developerslooking to add certain elements to their software. The cloud began as “infra-structure as a service,” then enabled “software as a service,” and has becomethe platform for hosting these reusable assets. But how do you know whichcloud to choose? As the market matures, decisions will have to be made.

Meanwhile, we’re seeing more open-source code being used in commer-cial software as developers realize they don’t have to reinvent the wheelevery time they want to create a program. But how and where open sourcecan be used remains a bone of contention in the industry, as different licens-es call for different use cases.

But atop of all of this, organizations realized the need for their systems tobe able to react to input and make recommendations or decisions on theirown. Therefore, the editors of SD Times have declared 2016 “The Year ofArtificial Intelligence.”

More than winning game shows or beating game champions, artificialintelligence represents the brains within these new systems, and 2016 saw anexplosion of APIs that enable developers to build such functionality as natu-ral language processing, personality insights and sentiment analysis into theirsoftware.

An exciting year, indeed. We invite you to enjoy our look back at the yearthat was on the following pages, and to share your thoughts with us via e-mailor comments online.

And, while we can’t imagine what 2017 will hold, we’re pretty certain many of you do. We’ve included “Predictions for 2017” from industry thought leaders to share what they’re thinking with you. z

SDT333 page 10-11_Layout 1 12/20/16 3:54 PM

BY CHRISTINA CARDOZA

Artificial intelligence isn’t a new con-cept. It is something that companiesand businesses have been trying toimplement (and something that societyhas feared) for decades. However, withall the recent advancements to democ-ratize artificial intelligence and use itfor good, almost every company startedto turn to this technology and tech-nique in 2016.

The year started with Facebook’sCEO Mark Zuckerberg announcing hisplan to build an artificially intelligentassistant to do everything from adjust-ing the temperature in his house tochecking up on his baby girl. Heworked throughout the year to bring hisplan to life, with an update in Augustthat stated he was almost ready to showoff his AI to the world.

In November, Facebook announcedit was beginning to focus on giving com-puters the ability to think, learn, planand reason like humans. In order tochange the negative stigma peopleassociate with AI, the company endedits year with the release of AI educa-tional videos designed to make thetechnology easier to understand.

Microsoft followed Facebook’s pur-suit of artificial intelligence, but insteadof building its own personal assistant,the company made strides to democra-tize AI. In January, the companyreleased its deep learning solution,Computational Network Toolkit(CNTK), on GitHub. Recently,Microsoft announced an update toCNTK with new Python and C++ pro-gramming language functionalities, aswell as reinforcement learning algo-rithm capabilities. In July, Microsoftalso open-sourced its Minecraft AI test-ing platform to provide developers witha test bed for their AI research.

But the company’s AI goals didn’tstop there. At its Ignite conference inSeptember, CEO Satya Nadellaannounced his company’s objective tomake AI easier to understand. “We wantto empower people with the tools of AIso they can build their own solutions,”he said. Following Nadella’s announce-ment, Microsoft formed an artificialintelligence division known as the “Part-nership on AI” with top tech companiessuch as Amazon, Facebook, GoogleDeepMind and IBM. Microsoft endedthe year teaming up with OpenAI to

advance AI research. Google started the year with a major

breakthrough in artificial intelligence.The company’s AI system, AlphaGo,was the first AI system to beat a masterat the ancient strategy game Go. InApril, the company announced it wasready for an AI-first world. “Over time,the computer itself—whatever its formfactor—will be an intelligent assistanthelping you through your day,” saidCEO Sundar Pichai. “We will movefrom mobile-first to an AI-first world.”

Pichai reiterated that sentiment atthe Google I/O developer conference inMay where he announced that the com-pany’s advances in machine learning andAI would bring new and better experi-ences to its users. For instance, thecompany announced the voice-basedhelper Google Assistant, updates to itsmachine learning toolkit TensorFlow,and the release of the Natural LanguageAPI and Cloud Speech API throughoutthe year. To help bring wider adoptionto AI, Google also created a site calledAI Experiments in November designedto make it easier for anyone to exploreAI. The year ended for Google with theopen-source release of its DeepMind


The year artificial intelligenceexploded

Companies made huge strides in using it for Big Data, making it easier to understand, and letting it watch the house

SDT333 page 12,14-17_Layout 1 12/20/16 3:46 PM

Lab, a 3D platform for agent-based AI research.

IBM, the company known forits cognitive system IBM Watson,also made waves in the AI worldthis year. The company started theyear with the release of IBM Pre-dictive Analytics, a service allow-ing developers to build machinelearning models. In October, thecompany announced the WatsonData Platform with MachineLearning, and a new AI Nanode-gree program with Udacity at itsWorld of Watson conference inOctober. The company ended theyear with the release of ProjectDataWorks, a solution designed tomake AI-powered decisions. Italso announced a partnership withTopcoder to bring AI capabilitiesto developers.

There was a smattering of AInews to be found as well. BaiduResearch’s Silicon Valley AI Labreleased code to advance speechrecognition at the beginning of theyear. NVIDIA began to developAI software to accelerate cancerresearch. Carnegie Mellon Uni-versity researchers announced afive-year research initiative toreverse-engineer the brain andexplore machine learning as wellas computer vision. Researchersfrom MIT’s Computer Scienceand Artificial Laboratory devel-oped a technique to understandhow and why AI machines makecertain decisions. Big Data com-panies turned to machine learningand deep learning techniques tohelp derive value from their data.OpenAI rounded out the yearwith the release of Universe, anew AI software platform for test-ing and evaluating the generalintelligence of AI.

Artificial intelligence is“intended to help people makebetter decisions. The systemlearns at scale, gets better throughexperience, and interacts withhumans in a more natural way,”said Jonas Nwuke, platform man-ager for IBM Watson. z


BY MADISON MOORE

In addition to the new testing suitesand solutions companies introducedthis past year, we saw the testing indus-try as a whole recognize that in order tokeep up with the pace of agile, theirown testing processes need to be incheck, especially if they want to stay ontrack with this new way of working.

The testing in production movementshifted its focus beyond just develop-ment; now it’s more about bringing per-formance and load testing into the pic-ture. Companies should realize testingin production is easy to add and hard tomess up, according to Antony Edwards,CTO of TestPlant. In 2017, softwareteams should consider the load they puton their servers, the variants that needto be tested, and the right tests that areneeded in general.

Testing in production takes effort,and so does software automation, whichthe industry has seen take over in termsof testing strategies. Businesses aremoving to faster releases, and time tomarket has changed. Last year, testingwas all about automation, and whilemanual testing is still relevant in anagile environment, developers areincorporating more automating testswith “test early, test often,” the oft-spo-ken mantra.

Testing shouldn’t stopyour agility, so compa-nies are going to haveto get used to digi-tized businesses,where software is thebusiness. Get yourteams “test-infected,”as Eric Taylor, director ofagile software delivery for Agile-Trailblazers, said, and devise a planwhere the process goes from business todevelopment to quality assurance.Teams need to change their mindsetand realize that testing should be first,as opposed to something that’s just

added at the end of the process, he said.That strong feedback loop and strongcommunication between the business,development and QA is the key to test-ing success.

This new mindset should be coupledwith testing automation. Companiesshould consider getting started withautomation by first figuring out theirneeds. Automation is not a one-size-fits-all solution.

For instance, one company might bebetter off with an open-source solution,whereas an off-the-shelf suite could bebetter for another. Rogue Wave prod-uct manager Walter Capitani said thatnew tools have contributed to the evo-lution of software testing, but “Thereality is any modern software that isbeing developed should have someform of automated testing.” If a compa-ny is not doing any sort of automation,they are really behind the times.

Companies are seeing the value oftesting early and often, but there werestill too many vulnerabilities in softwaresecurity efforts. We saw cybersecurityremained a huge problem in 2016’sworld of software testing. Too manycompanies are relying on bug bountyprograms, which only provide a quickfix (and can be expensive), according to

a September report released by Vera-code.

The same report said 83% ofIT decision-makers admitted to releasing code before testingor resolving issues. ChrisWyspoal, CTO and cofounder

of Veracode, said that securityneeds to be a part of the entire

software life cycle. The best way toget IT teams to perform adequate

security testing is to listen to cus-tomers, train developers,

and incorporate threatmodeling into thedesign process,he said. z

Testing kept upwith agile in 2016



BY ALEX HANDY

Last year, the world was excited aboutthe prospect of Docker lowering cloudcomplexity and removing the need torun virtual machines in the cloud. Thisyear, however, the industry seemed toawaken to the need to be cross-cloud-capable to avoid vendor lock-in.

That industry-wide realization cameat a time when Microsoft’s Azure wasturning into not just an Amazon alter-native, but a legitimate competitor.The company doubled its revenues onAzure this past year, but that doesn’tmean it was slowing down. In fact,Microsoft spent a billion dollars in2016 alone on data centers in Europeto expand its cloud operations there.

Which isn’t to say Amazon is not stillthe leader in the cloud. That company’sweb service offerings continued toexpand in 2016. Among the new fea-tures were HIPAA compliance formore of its services, price reductions,and new features for Storage Gateway.And at the end of the year, support fordesktop Windows 10 deployments.That last addition gave users access todesktop computer use from the cloud.

Amid all the wild news fromMicrosoft and Amazon, you’d expectGoogle to be racing alongside with itsown big announcements. But Googlespent most of 2016 preparing for arevising of sorts of the Cloud Platformas a whole. After acquiring her startupin 2015, Google installed VMwarefounder and former CEO DianeGreene as vice president of Google’scloud business.

This hire was supplemented later inthe year with the acquisition of Apigeeand the hiring of former Apigee strate-gy chief Sam Ramji. The underlyingshift in the platform may not actuallybe a shift in the platform itself, but

rather is a major shift in the strategybehind bringing Google Cloud to mar-ket. The company has already mademoves to keep the platform competi-tive, and as of this writing, GoogleCloud is about half the price of Ama-zon for certain services. But we expect2017 to be the year the dial begins tomove on the somewhat left-behindGoogle Cloud.

No matter how Google does in2017, it’s a safe bet that OpenStackcontinues to grow and mature. Thisopen platform for data centers hasgrown significantly over the past fewyears. Today, it’s got an appealing posi-tion inside the enterprise as a bridgebetween the public and private clouds.

Interestingly enough, OpenStack’soriginal value proposition of offeringinternal developers the same on-demand cloud-like hosting capabilitiesavailable in AWS and Azure has nowgiven way to a much more mundaneuse case: maintaining infrastructureportability.

Developers are now using Open-Stack as a way to push their internaldeployments, and are then bursting toexternal clouds when demand peaks.This ensures applications cannot sim-ply be written to work in Amazon andnowhere else.

And that’s a major trend we expectto see continue in 2017: avoiding ven-dor lock-in. While Amazon has provid-ed developers with some of the easiestand most powerful APIs and servicesinside its cloud, using them is tanta-mount to pouring concrete on yourapplication’s feet and shoving it intothe ocean: Once you’ve built with aspecific cloud’s services, it’s unlikelyyou’ll be getting away from them anytime soon.

That’s the one thing every softwaredeveloper and manager learned toavoid back in the 1990s. Only now isthe industry starting to see it is possibleto build multi-cloud applications with-out having to invest in significant inter-nal development to make it happen. z

The cloud matures, Amazonsees stiffer competition While Microsoft forges ahead on Azure, Google bides its timefor 2017, and everyone agrees that vendor lock-in is bad


SDT333 Full Page Ads_Layout 1 12/20/16 12:17 PM Page 13

http://www.cdata.com


BY ALEX HANDY

Depending on where you stand, 2016was either the best year ever for open-source software, or it was a year of con-troversy and danger. While it’s undeni-able that 2016 saw more contributors toopen source and more open-sourceprojects than any prior year, it’s also truethat this was a year of strife for commu-nities, developers and users alike.

Chief among those problems wouldhave to be the Dirty COW local privi-lege escalation attack, a major vulnera-bility that seems to have been hidinginside the Linux kernel for the pastnine years. The discovery of this exploitisn’t necessarily a knock against open-

source software as a whole: The bugmight never have been found if thesources weren’t also available.

The Dirty COW bug, however, high-lighted one of the most difficult prob-lems challenging open-source softwareecosystems, and one that in particularaffects Android. How does one evenbegin to comprehend how to patch thishole across the Android ecosystem? Withbillions of Android phones out there, anda huge number of them no longer sup-ported or updated by their manufactur-ers, it would appear that the Dirty COWwill be a going concern in Android untilthe heat death of the universe.

This is just another part of the per-

petual fragmentation problems Androidhas had to deal with from day one. It’s adramatic difference from the world ofservers and desktops, where patcheswere issued within a month of discovery.Here, Linux has solidified into a reliableplatform with long-term support vectorsand emergency patch routines andchannels. But the Android ecosystemdoesn’t even have a concept of long-term support.

Elsewhere in the open-source world,the Apache Foundation pushedJonathan Ellis out of his position as chairof the Apache Cassandra Project. WhileEllis’ involvement with the project willin no way decrease, the move was indica-

BY MADISON MOORE

DevOps is the buzzword software and technology compa-nies are all too familiar with. Despite how mainstream theconcept is, organizations are only just now implementingappropriate DevOps strategies, especially since automation,continuous testing and getting to market faster are the ways

to remain on top.DevOps is designed to improve the quality of applica-

tions, but it’s also designed to improve the quality of theorganization itself. This year, organizations explored newways to get their teams to adopt DevOps, like making suremanagement plays a new and different role, and truly break-ing down silos. These are steps that vice president of com-munity development at Chef, Nathen Harvey, recommendsteams consider in the future.

Another way companies can break down siloes is to alignthe people with the processes and technology. The mainissue agile experts like Zubin Irani, CEO of cPrime, saw withorganizations is that they approach agile from differentangles, and as a result, there is absolutely no connectivity.Agile and DevOps are meant to solve problems, but theseproblems cannot be solved unless companies see the big pic-ture, he said.

In addition to companies ramping up their collaborationefforts, there were a few big announcements in terms ofDevOps solutions and testing suites. CA Technologies

Open source grows, but conflict remains

Microsoft firmly embraces

it, while the Linux

Foundation grapples with

commercial concerns

Companies are reaping the Mainstream adoption means new tools and a growing need for changes in practices



acquired BlazeMeter in October inorder to strengthen its position in test-ing efficiency and to broaden itsDevOps portfolio.

CA later acquired Automic inDecember after releasing its unifiedsuite of business automation productsfor DevOps teams a few months earlier.The acquisition of Automic would givecustomers more options to addresstheir IT and DevOps challenges, espe-cially for on-premise, the cloud andhybrid cloud environments, said AymanSayed, president and chief product offi-cer of CA.

CollabNet also joined the DevOpsmovement this year, after it introduceda DevOps solution and partnershipwith Clarive Software. Its DevOps Life-cycle Manager was created to helpteams deliver applications faster whilestill having visibility across the entiretool chain of work items.

HashiCorp ventured into enterprisesales this year with Vault Enterprise, itsDevOps tool chain that enables organi-

zations to adopt DevOps practices intheir approach to security. Specifically,Vault offered developers a simplerinterface for dealing with hardwaresecurity modules, along with a unifiedpath for developing security modules.

Tuleap wanted agile teams toimprove their engineering practiceslike code reviews, Continuous Integra-tion and regular release delivery. Its lat-est release, the open-source Tuleap 9,came with an easy-to-use kanban tooland production-scale Git for DevOpsteams.

Security was also a big topic of con-versation this year for DevOps teams,because the shift in speed and agilityexposed some major gaps in applicationsecurity, according to an HPE SecurityFortify survey of security developers.

According to this report, 99% ofrespondents agreed that DevOps cul-ture needs to improve when it comes toapplication security, and yet only 20%actually test security throughout theirdevelopment processes. A lack of secu-

rity testing in DevOps environmentscreates some barriers for teams, andmoving forward, companies shouldconsider the top-down and bottom-upapproach when it comes to security.Security should be a shared responsibil-ity, and developers need to help identi-fy these vulnerabilities earlier on in thedevelopment life cycle.

In the new year, it’s possible DevOpswon’t be around anymore. Not thepractice or processes, but the worditself. According to Todd DeLaughter,CEO of Automic Software (now a partof CA Technologies), today’s catchytechnology phrases have become “sointrinsic” to business and operationsthat there isn’t a need for them any-more. And DevOps is at the top of thelist of words that are no longer neededsince it’s becoming more mainstream.

DevOps is on its way to achieving itsfull potential, but the only way to knowwill be if people stop saying its name,said DeLaughter. Perhaps “DevOps”will be dead in 2017. z

tive of a larger concern within the Foun-dation over increasingly blurry linesbetween commercial enterprise soft-ware offerings and their open-source,free counterparts under its jurisdiction.

Even the Free Software Foundationsaw some waves of controversy thisyear, with its general counsel EbenMoglen stepping down on Oct. 27 andleaving a vacancy still unfilled in theposition. While he did not respond to

requests for comment, mul-tiple sources have intimated

that his departure cameat the behest of

Richard Stallman himself, who feltMoglen was no longer in sync with theFoundation and movement. Moglenhad been the FSF’s general counsel formore than 20 years.

Fortunately, 2016 also saw somemajor updates to open-source projects.Open-source efforts in the JavaScriptworld, such as Angular 2, Node.js andMeteor, all grew and expanded their

feature sets in 2016.Yet in keeping with the theme of

controversy in 2016, Node.js hadits own problems. When left-pad,

an 11-line bit of code in theNPM repositories, was

removed due to its devel-oper becoming irate, it

broke thousands ofapplications around theworld, some of themmission-critical.

For 2017, you canexpect the Node.js and

NPM community to continue to mopup after this embarrassing failure. Infact, NodeSource has already taken upthe reins by offering an enterprise-focused service that blesses NPM pack-ages as worthy of corporate use.

Still, no amount of controversy orsuccess could cover up the biggest open-source news of 2016: Microsoft. In thepast, Microsoft had been outright hostileto open source. Even its attempts tomake up with the open-source worldwere ham-fisted, like the company’sCodePlex efforts in 2008, which createdwhat it called an “open-source museum”of look-but-don’t-touch software.

Today, however, Microsoft is a verita-ble pillar of the open-source community.Whether it’s supporting Linux on Azure,or building the next revisions of C# andthe .NET platform in the open-sourcecommunity, Microsoft’s movement toopen source has to be the biggest andmost dramatic story of the year. z

benefits of agile, DevOps


John Schroeder, executive chairman and

founder, MapR

AI is now back in mainstream discus-sions, and is the umbrella buzzword formachine intelligence, machine learning,neural networks, and cognitive comput-ing. Why is AI a rejuvenated trend? Thethree V’s come to mind: Velocity, Varietyand Volume. Platforms that can processthe three V’s with modern and tradition-al processing models that scale horizon-tally providing 10x-20x cost efficiencyover traditional platforms.

Google has documented how simplealgorithms executed frequently againstlarge datasets yield better results thanother approaches using smaller sets.We’ll see the highest value from apply-ing AI to high-volume repetitive taskswhere consistency is more effectivethan gaining human intuitive oversightat the expense of human error and cost.

Eric Mizell, vice president of global

solutions engineering, Kinetica

Real change is coming to real-timeintelligence in 2017 with graphical pro-cessing units. GPUs are capable ofdelivering up to 100x better perform-ance than even the most advanced in-memory databases that use CPUs alone.The reason is their massively parallelprocessing, with some GPUs containingover 4,000 cores, compared to the 16 to32 cores typical in today’s most powerfulCPUs. Amazon has already begundeploying GPUs, and Microsoft andGoogle have announced plans. Thesecloud service providers are all deployingGPUs for the same reason: to gain acompetitive advantage.

Viktor Farcic, senior consultant,

CloudBees

A significant change in 2017 will befocused not so much around runningcontainers, but scheduling them insideclusters. Solutions like Docker Swarm,Kubernetes, Mesos, etc. will becomemainstream. We will see more solutionsthat will go beyond simple scheduling.We’ll see the rise of self-healing sys-tems. In 2017, the battle for the “uberorchestrator” will become much moreprominent.

One of the most exciting areas that

will become prominent in 2017 will beunikernels. While the majority of theindustry is still trying to wrap theirheads around containers, we will startseeing unikernels taking over the stage.They will, in a way, unify functionalitiesprovided by VMs and containers.

Abdul Razack, SVP and head of platforms,

Infosys

Despite the advancements we’ve madein artificial intelligence, it is not a one-to-one replacement for people. Whilethe technology amplifies human abili-ties, you cannot teach AI systems to beproactive, creative or think on theirfeet. In 2016, AI was applied to solveknown problems. And as we move for-ward, we will start leveraging AI to gaingreater insights into ongoing problemsthat we didn’t even know existed. UsingAI to uncover these “unknownunknowns” will free us to collaboratemore and tackle new, interesting andlife-changing challenges.

Scott Miles, senior director of cloud,

enterprise and security portfolio market-

ing, Juniper Networks

In 2017, automation and analytics willhelp organizations address the shortageof security personnel. Often organiza-tions invest heavily in effective security

hardware and software, but lack thesecurity specialists necessary to ensuretheir effectiveness.

As an example, breaches like the onesthat impacted Target and Home Depotwere detected by their high-end securitysystems, but the security operationspractitioners were too overwhelmed bythe thousands of alerts they received perhour to see which ones posed the mostimminent threat. As automationbecomes more integrated into securitysolutions, security personnel will receivefewer notifications with more relevance,relieving them of the manual task ofhunting through a sea of alerts to findthe truly malicious ones.

Roy Solomon, cofounder and vice presi-

dent of product management, Applause

Companies of all sizes will put moreemphasis on great digital experiences asusers expect brands to engage withthem through the digital channel oftheir choice, not the other way around.To facilitate this, companies will needto ensure all of their digital platformscreate seamless user interactions,which means a heavier focus on techni-cal elements that go beyond traditionalQA such as APIs, in-app bug reportingand crash reports, as well as on greatusability.

SD Times January 201718

Experts see more changes in DevOps, and how developers interact with

What’s on the

SDT333 page 18,19NEW_Layout 1 12/21/16 1:45 PM

David Lee, vice president of platform

products at RingCentral

There will be an exponential increasein the use of Web Real TimeCommunication (WebRTC) servicesfor web and mobile apps. Driving thisadoption is the enterprise need to pro-vide rich interactions where workerscan communicate seamlessly with eachother. By being able to build thosecapabilities out of HTML andJavaScript, WebRTC will increasinglybecome the easiest and most popularchoice for developers to embed real-time communications capabilities intocustom applications.

Andrew Levy, Co-founder, Apteligent

2017 is a critical year for Microsoft. Theyhave failed in the phone market with lessthan 1% market share, took a $7.6 billionwrite-down on their Nokia acquisition,and laid off thousands of employees. Italso looks like they will discontinue theLumia brand. On the flipside, they havefound success with their Surface brand,especially the Surface Pro.

Most expect Microsoft to introducea Surface Phone in 2017. Its Continu-um product makes this especially inter-esting since your phone becomes yourmobile desktop, capable of connectingto a larger monitor and input deviceslike a mouse and keyboard. It’s not clear

where Microsoft goes in the smart-phone market from here if the SurfacePhone fails. Perhaps it’s their invest-ments in the HoloLens. If this type ofwearable is the future of computing(and telecommunications), then theywill be well positioned for the future.

Ashley Kramer, director of product

man agement and head of cloud strategy,

Tableau

Many organizations are living a hybridreality split between on-premises andcloud environments—cloud is no longerisolated from your on-premises data andinfrastructure. Vendor investments inmigration tools and strategies will helpcustomers navigate through this hybridworld. For the end user, these solutionsmake complex hybrid environmentsfunction as one cohesive system. Invest-ments in hybrid software will remainfully relevant even as organizations shiftoperations toward an all-cloud future.

Sean Regan, JIRA Software and

Bitbucket team lead, Atlassian

The small, agile software team has nev-er been more empowered as microser-vices, Continuous Delivery and cloudplatforms rise in prominence and accel-erating software development. Thebottle neck today isn’t hardware; it is

people. You can’t scale people like youcan scale cloud computing.

As a result teams will be looking forways to collaborate better and faster onhigh-value work and less on low-valuerepetitive work. We can expect to see asignificant investment in collaborationand automation tools for DevOps prac-tices, where simplistic tasks that requirethe speed and efficiency of moderncomputing will be handed over to the“robots” to accomplish.

Sirish Raghuram, CEO, Platform9

We will see serverless technologiesgain real traction in 2017. The first usecase that will drive adoption in 2017 isbot development. Today, a lot ofDevOps automation involves writingbots that integrate with various systemsusing Web Hooks. Serverless makesthis incredibly natural and easy. Anoth-er major use case is making it easierand faster to start using containerswithout having to fully learn andunderstand all the concepts to systemssuch as Kubernetes.

With an intuitive experience thatabstracts away the underlying layers ofclusters, pods, networks and storage,serverless presents a developer-friendlyconsumption paradigm for containersand Kubernetes.

Jeff Prus, vice president of product management at QuickBase

Developers’ roles will be redefined:They’ll no longer be considered “IT”workers. In 2017, we’ll see organiza-tions start thinking outside the box tohelp fill development needs and seekout empowered problem solvers,regardless of where they sit in theorganization. While coding skills willcontinue to be important and in highdemand, in many cases they will nolonger be the end-all, be-all.

While the definition of “developer”won’t change overnight, 2017 is likely tomark the onset: Gartner predicts thatby 2020, 60% of all fast-mode applica-tion development projects will be doneoutside of formal IT teams. This shiftwill present new challenges, as organi-zations of every size change the waythey attract, evaluate and measure theperformance of developers. z


, in AI, in what it means to be in IT, h each other and their environments

horizon for

SDT333 page 18,19NEW_Layout 1 12/21/16 1:45 PM

C

M

Y

CM

MY

CY

CMY

K

Big Data Whitepaper ad SD TIMES-1.pdf 5 24-10-2016 2.01.18 PM

http://www.syncfusion.com/SDbigdata


BY MADISON MOORE

During the holiday season, retailerswatched their online sales climb. Butwhile businesses may have had awindfall, consumers were likelyunaware that 100% of those retailershad issues with domain security,according to SecurityScorecard’s“2016 Biggest Holiday RetailersCybersecurity Report.”

When it comes to cybersecurity,retailers have a lot to worry about,especially since they handle billions oftransactions each year. All of that con-sumer data needs to be protected, andoftentimes, according tothe report, it is the largestretailers that “succumb todata breaches.”

In February, NeimanMarcus also suffered fromits own data breach, when5,200 customer accountswere accessed throughautomated attacks.

SecurityScorecard want-ed to see if other retailers were at riskfor data breaches and other securityvulnerabilities, so between April 1 andOct. 31 of last year, its securityresearchers looked at the 48 biggestretailers that collectively representmore than US$1 trillion in annual sales,according to the report. Some of theseretailers included Amazon, Costco,Lowes, Macy’s, Sears, Staples, Targetand Walmart.

According to Alex Heid, chiefresearch officer at SecurityScorecard,the “correlation between disclosedbreaches, leaking credentials, and hack-er chatter” were some of the interestingfinds from their research.

“In the retail industry, it seems thatthe circulation of compromised creden-tials and shared fraud methodologiesare more open and public as comparedto the financial industry or insuranceindustry,” he said. “In financial andinsurance, admissions of breaches are

oftentimes attempted to be obfuscatedor hidden in an effort to ensure longevi-ty of the looted information.”

Everybody’s got problemsA big finding from SecurityScorecard’sreport was that all of the 48 retailers ana-lyzed were found to have multiple issueswith domain security, which indicatesthat retailers’ domains aren’t configuredproperly to defend against hackers orimpersonation attacks.

Additionally, nearly 80% of retailersmay not be using intrusion detection orprevention systems to monitor their

traffic within the cardhold-er data environment. Heidsaid that many hackers hitretailers in the form of webapplication attacks, and theuse of web application fire-wall technologies aids inthe detection and mitiga-tion of common web appli-cation attacks, he said.

SecurityScorecard alsolooked at the overall performance ofcompanies over a seven-month period.According to the report, these retailershave been struggling with maintaining ahigh grade in the problematic securitycategories, such as network security,where 69% of retailers had multipleentry points for hackers. Also, 73% ofretailers had misconfigured websitedomains, which impact DNS health.

“With grades like these, if a hackerdecides to take action while organiza-tions scramble to keep up with anuptick in activity, they may find an easyway into the organization’s network,opening the floodgates to a potentialdata breach,” read the report.

Another analysis in the report exam-ined the Payment Card Industry (PCI)Data Security Standard, which is a set ofsecurity standards for any retailer thatprocesses, stores, or transmits creditcard information. The fines for compa-nies that do not comply with these stan-

dards could cost companies tens ofthousands of dollars, said the report.

SecurityScorecard looked at howwell the 48 retailers met the PCI com-pliance standards to get more insighton possible shortcomings, and thecompany found that 50% had issuetypes that “may be grounds for a com-pany failing to meet the standard,” saidthe report.

Part of the reason why these retail-ers are failing to meet security stan-dards is due to their large size, accord-ing to SecurityScorecard in the report.

“Zero-day vulnerabilities that aren’tpatched, low security awareness amongemployees, weak network security, andimproper domain and e-mail configura-tions are all signs of slow-moving andinflexible companies that aren’t quick toreact to potential new risks,” the reportread. “Given the influx of new activitythat these companies need to considerwith the upcoming holiday season, extraresources and attention should be leviedon their security department.”

Moving forward, SecurityScorecardrecommends better security awarenesstraining for employees of these largeretail companies. With adequate train-ing, employees may be able to “fend off”phishing and other vulnerabilities thatcan take over their systems. Companiesalso need to make sure they protecttheir internal and external customers,especially private and sensitive informa-tion like passwords and credit cardinformation.

“Retailers should examine the use ofnext-generation firewalls, endpointprotection solutions, and web applica-tion firewall technologies in combina-tion with a continuous informationsecurity monitoring solution that exam-ines both the security posture of theenterprise, as well that of their relatedpartners and vendors,” said Heid.“Many breaches originate from inse-cure third parties that provide a path-way to exploitation.” z

Top retailers are open to hacks SecurityScorecard found that domain security was a universal problem

SDT333 page 20_Layout 1 12/20/16 12:06 PM Page 20


http://www.nevron.com


http://www.xceed.com

BY ALEX HANDY

At AnDevCon in December, TimothyJordan, head of Google platform devel-oper relations, gave a keynote to detailthe current and future state of the plat-form and its ensuing toolsets. Hedescribed changes coming in Android7.1, and he taught attendees how to getstarted with TensorFlow on Androiddevices.

The next release of the Android plat-form will offer developers new featuresdesigned to make their applicationseasier to use. First in line is app short-cuts, which would allow users to selectfrom a set of predefined actions whentapping an app icon.

With an app shortcut, users long-pressing a messaging application mightbe offered a pop-up list of tasks that canbe performed within it. These couldinclude creating a new message orjumping to a recent conversation.

These new app shortcuts aredescribed with CML files, and develop-ers can build these actions aroundintents. Jordan said that this will enableusers to get to the functionality they needquickly and easily from the home screen.

Android 7.1 will also add support forround icons, which can be created in theImage Asset Studio. Another image-based feature coming in 7.1 is support forimage-based keyboards. For usersinputting text into an application, thisnew Commit Content API will allowimages and other rich content to bepushed into the text field.

Applications will declare what types

of media they support, and doing so willallow users to select images or other con-tent from a sliding bar at the bottom ofthe screen, acting as a sort of keyboard.The new API is also supported all theway back to the Honeycomb releases ofAndroid.

Retailers will most be able to makeuse of Demo User mode. This enablesdevelopers to throw their applicationsinto demo mode, making them performin a manner more appropriate to beingshown on a retail floor. Some examples ofdemo mode toggles include disabling thecreation of new users, or removingbilling screens.

Jordan also discussed forthcomingchanges to Android Studio 2.3. Theseinclude an overall update to the latestversion of IntelliJ, as well as new supportfor lint checks. This version also addssupport for Android 7.1.

As for Firebase, Jordan said it hasexpanded its offerings to support devel-opers on the back end of their applica-tions. New features for Firebase includeUnity support, better analytics, and anew Udacity course to teach developershow to use the service.

Jordan spent half his time talkingabout these new products and features,and then started a half-hour tutorial onhow to build a Hello World project withTensorFlow, Google’s machine learninglibrary.

“When I talk to developers world-wide who are building applications andthinking about machine learning, itsounds very complicated the way they

describe it. And it is,” said Jordan.“However, the first overview is simplerthan you think. You can start under-standing it pretty quickly, and you canstart using it even quicker than that. I’mnot an expert in machine learning, but Iam an expert in the tools. That’s what’sexciting now. It’s the first time we’vebeen able to access this level of intelli-gence without having two or threePh.D.’s on our team.”

Jordan then described some ofGoogle’s machine learning offerings.On mobile devices, for example,Google’s Vision API can be run locally.He also discussed Cloud MachineLearning support in Google Cloud,which can train machine learning algo-rithms in Google’s data centers.Google’s Vision API is open source andavailable for free.

Jordan demonstrated some simplemachine learning capabilities, such asstyle transfer on artworks. Using styletransfer, the developer can combine twoimages, merging their imagery and styletogether. TensorFlow was used to buildGoogle’s Deep Dream project, whichcreated truly bizarre images by analyzingand reimagining them with a machinelearning algorithm.

TensorFlow can be used to do textu-al analysis through Parsey McParseface,an open-source library released earlierthis year, according to Jordan. He saidthat the work Google is doing onmachine learning is no longer in theresearch phase, and is now ready forbusiness use. z


Phot

o by

Mar

k Sc

heel

Timothy Jordan gave both a lecture on Android 7.1 and lessons on TensorFlow at AnDevCon.

Android 7.1: What you can expectApp shortcuts, round icons, and expanded Firebase offerings coming in 2017

SDT333 page 23_Layout 1 12/20/16 11:59 AM Page 23


BY ALEX HANDY

Mitchell Hashimoto has been writingsoftware since he was 12 years old.Since he cofounded HashiCorp in2012, however, he’s been focused onautomation software, such as that creat-ed by his company. HashiCorp nowoffers a host of products to automatethe software development life cycle,and those products do everything frommanaging security between services toprovisioning environments in the cloud.

We spoke with Hashimoto about hiscompany’s past, present and future, andto discuss the impact of Vagrant on hislife and work.

SD Times: HashiCorp’s tools require acertain way of thinking about softwaredevelopment and deployment. Can youdescribe that?

Hashimoto: We have something wepublished called the Tao of HashiCorp.One of our core beliefs is how we thinksoftware and infrastructure automationshould work. Those tenets are infra-structure as code, and really everythingas code.

We believe human memory is falli-ble, and you want a source of truth foraccountability and history. The onlyway to get automation on top of that isto write something down in a very liter-al way, which is code.

We favor declarative versus impera-tive. We believe everything should bedeclarative at a certain scale and com-plexity. If you’re writing down how todo something, it becomes too complexversus saying what you want to happenand letting things get there.

It’s scarier because you have a lack ofcontrol getting to your desired state,but as something matures, when youuse a declarative system really well, it’sindistinguishable from magic.

From an organizational perspective,you should be striving to automate asmuch as possible. When I look at people,I think what makes us special is beingcreative and creative problem solving:adapting to things we haven’t seenbefore. Whenever I found myself doingsomething that wasn’t creative, like arepetitive task, it felt like a waste to me.

The gift we have is this creativity, andit’s great when that’s focused on softwareor marketing or whatever. We should bestriving to automate as many rote tasks aspossible so we can do creative things.

This all got started with Vagrant. Whatdoes that tool mean to you now?

Vagrant in a lot of ways is the first thingI think anybody does. There’s a lot oftimes where I didn’t realize the deci-sions I were making, but Vagrant caughta lot of them. We solved a smaller-scopeproblem really well, versus trying to doeverything at once. It also embodied theproblem set I described. I found myselfdoing this rote task whenever I hoppedinto new client work. I had to re-set upmy laptop, reading a readme [and]doing what the readme said. I looked atit and said there’s no reason a humanshould do this.

Did developing Vagrant awaken you tothe need for more automation toolingin the life cycle?

It was all kind of in parallel. When I was

working on Vagrant with Armon Dadgar,my cofounder, we were working on alarge-scale research system. We had allthese other problems. In hindsight, wewere experiencing the pains of microser-vices and multi-cloud and problems inthat space, and we just couldn’t managethat scale for that application deliverylife cycle. We couldn’t manage it withthe tools we had. We were pretty young.

We were writing down the chal-lenges we were facing. As Vagrant grewin popularity and we got more involvedin the DevOps movement, I startedseeing that this movement, these prob-lems and this academic research all lineup to solve these problems.

What made you build out more toolsand found HashiCorp?

Vagrant hit a limit. How do I actuallydeploy this thing? And there was nogood answer. We started looking atwhat additional software we couldmake. When I announced HashiCorp,we came out with Packer, but over atwo-year period we released a lot.

It was more intentional. When yourelease a puzzle piece and you haven’tgiven out all the pieces, it adds moreconfusion. There were a number ofpieces we wanted to get out there, so itwas a semblance of a puzzle. We need-ed the first five to be out there to give adecent end-to-end story: to give a majorkey frame story for getting from devel-opment to production.What’s next for HashiCorp?

In the future, there are a few trends wefind interesting. One in particular we’re


HashiCorp and thestate of automationMitchell Hashimoto talks about what to expect for deployment in 2017

SDT333 page 24,26_Layout 1 12/20/16 12:05 PM Page 24


http://www.aspose.com


BY CHRISTINA CARDOZA

The Node.js Foundation is continuingits mission to make Node.js VM-neu-tral. The foundation announced majormilestones toward allowing the solutionto work in a wide variety of VMs at theLinux Foundation’s Node.js Interactiveconference.

According to the foundation, VM-neutrality will allow Node.js to expandits ecosystem to more devices andworkloads, such as the Internet ofThings and mobile devices. Other ben-efits include developer productivity andstandardized efforts.

As part of VM-neutrality, the founda-tion has announced that the Node.js APIis now independent from any changes inV8, the open-source JavaScript engine.“A large part of the Foundation’s work isfocused on improving versatility andconfidence in Node.js,” said MikealRogers, community manager of theNode.js Foundation. “Node.js APIefforts support our mission of spreadingNode.js to as many different environ-ments as possible. This is the beginningof a big community web project that willgive VMs the same type of competition

and innovation that you see within thebrowser space.”

In addition, the foundationrevealed the Node.js build system willstart to produce nightly builds of node-chakracore, allowing Node.js to beused with Microsoft’s JavaScriptengine, ChakraCore.

“Today, there is a proliferation in thevariety of device types, each with differ-ing resource constraints,” wroteArunesh Chandra, senior programmanager for Chakra, in a blog post. “Inthis device context, we believe thatenabling VM-neutrality in Node.js andproviding choice to developers acrossvarious device types and constraints arekey steps to help the Node.js ecosystemcontinue to grow.”

The Node.js Foundation alsoannounced plans to oversee a Node.jssecurity project at the conference,which is designed to detect and disclosesecurity vulnerabilities in Node.js.According to Rogers, the foundationwill allow security vendors to contributeto its common vulnerability repository.

“Given the maturity of Node.js andhow widely used it is in enterprise envi-

ronments, it makes sense to tackle thisendeavor under open governance facil-itated by the Node.js Foundation,” saidRogers. “This allows for more collabo-ration and communication within thebroad community of developers andend users, ensuring the stability andlongevity of the large, continuallygrowing Node.js ecosystem.” A Node.jssecurity project working group will beestablished as part of the Node.jsFoundation.

In other Node.js news, enterpriseNode company NodeSource announcedit is expanding its production toolset withNodeSource Certified Modules and therelease of NSolid v2.0. NodeSource Cer-tified Modules is designed to providesecurity and trust to third-partyJavaScript solutions. The solution verifiestrustworthiness through the NodeSourceCertification Process, and it ensures astable, reliable and secure source.

NSolid v2.0 is the latest release ofthe company’s enterprise-grade Node.jsplatform, and it features automatederror reporting, real-time metrics,built-in security features, CPU profil-ing, and performance monitoring. z

Notes from Node.js InteractiveVM neutrality, security project, and NodeSource NSolid 2.0

really latching onto is shifting this infra-structure as code to more categories.Our two biggest growth drivers in thepast 18 months have been Terraformand Vault. Terraform is for creatinginfrastructure as code, and Vault is oursecurity tool that provides certificatemanagement and key management.They are both seeing the same level ofgrowth: multi-100% growth every quar-ter. They are related in where they areheading, in that our road map revolvesaround pushing those even further.

Let’s describe more things as code.But also, let’s bring security as code intoVault. Right now, Vault is very impera-tive. It’s, “This person can access this

secret”; it’s very much old school. Youhave to do that with security so theyfeel comfortable.

As we gain heavy adoption of Vault,we’re starting to look into what is thenext step in security. How do we jumpsecurity to make it more manageable.There’s a real problem we’re seeing inmicroservices. It’s so difficult to reasonabout security. You’re in the cloud,there are no end points or out points;it’s just sort of the Internet. What youreally want is to secure every connec-tion with every service really fast.

We think that when we see thatcomplexity, when I look at it, you needa declarative system. You need to trustthe security systems, and that’s a big

leap. But I don’t see a future where youhave a thousand microservices andyou’re connecting every tube together.I don’t see a future where a securityengineer is reasonably doing that. Youneed more automation there.

What do you find interesting in tech-nology these days?

I am paying attention to the serverlessstuff popping up. None of it says to methis is the way to do it, but I do think if Ihad a lot more free time, I would be play-ing a lot more with serverless. It’s reallyunknown what the business value is, butthat unknown is really intriguing to me.Anything that pops us as serverless, I takea look at it. It can be fun, but I see a lotof challenges scaling that up to real busi-ness uses, like business analytics. But Ithink there’s something there. z

The state of automation< continued from page 24



http://www.groupdocs.com


http://www.nsoftware.com


BY NICK BITZER

On the day of the Software TestingWorld Cup Finals in early December, itwas an understatement to say my teamand I were nervous. We knew we hadworked our hardest to prepare, but thefact that none of us knew exactly what lieahead for us to test left us with an air ofuncertainty. The stakes were high on thebiggest global stage for software testing,and the pressure was rapidly mounting.

We started off the day by gettingphysically set up for the competition,held in the largest conference room atthe Dorint Hotel in Potsdam, Germany.Each of the seven teams had their ownworkstation, although they only consist-ed of a couple of tables along with somepower supplies. We were also givenlarge paper flip boards for whiteboard-ing ideas or problem-solving while inthe heat of the battle. Our laptops werecharged and primed to operate at peakspeed and performance, and our phoneswere juiced and ready.

Once we had our workstationprepped, we headed off to a quick inter-view with an STWC judge who quizzedus on how we prepared for the competi-tion and how we planned to go abouttesting the still-unknown application. Wetried not to give away too many secrets.

After the interview, we retreated toour workstation. As in the prelims, eachteam member was assigned an area thatthey would focus on testing. Whereas Ihandled security in the previous round,my area of concentration changed tobecome application performance and

exploratory testing. When you’re test-ing in that type of team and high-pres-sure environment, the key is to remainflexible. The rest of the team waspoised to focus on everything fromsecurity to third-party application inte-gration to static code analysis.

In order to be sure that we wereready to hit the ground running whenthe timer started, we had every testingtool we needed open and running. Thisincluded things like Android Studio forapp debugging, and Slack so that wecould communicate important facts orideas without other teams being able tooverhear us. (The workstations in theroom were only about 10 feet apart andhad no sound or vision barriers, addinganother wrinkle to the challenge.)

It was also important that I spentsome time double-checking that all thedeveloper options I needed wereenabled on both of the Android devices Iwas set to use: an LG Nexus 5 runningAndroid 4.4.4 (KitKat), and a Dell Venue8 tablet running Android 5.0 (Lollipop).Other members of the team strategicallyutilized different brands and types of

devices along with various versions ofAndroid to ensure that we had compre-hensive OS and hardware coverage whenit came time to test the application. Thatturned out to be a smart move.

The game beginsNo more than 30 minutes before thecompetition started, we received an e-mail informing us that we’d be testingan Android app called Moovel, a transitapplication that enables consumers tobook public transit or ridesharingoptions on the go. The e-mail containedinformation from the Moovel productowner about what aspects of the appthey were most concerned with testing,hardware/software requirements, and adebug .apk file, as well as what was inand out of scope for testing.

Once we got the application, wewere able to divvy up specific objectivesfor testing and estimating our testingprocess. I’ll admit it was tough to layout an official test plan. When you don’tknow what kind of application you’regoing to be testing, and you’ve got sucha finite timeframe, exploratory testingtends to be the best way to find bugs.

Upon seeing the link to the debug.apk, a wave of relief washed over me.Since I had a debug version of the app, Iknew I could do a lot more with perform-ance metrics in Android Studio. With thedebug application, I’d be able to gatherCPU, memory and network performancestatistics in a more reliable fashion.

It was finally our moment of truth:time to start testing. We actually had twohours and 40 minutes to test the app,write bugs and put together a finalreport, which was abbreviated from an


Nick Bitzer is a DevOps engineer at ReadyTalk.

Ready, set, TEST!Competition helped one team to get a sense of scope, asssigning objectives on the fly

Team “RT Pest Control” (clockwise from

left): Samantha Yacobucci, Zala Habibi,

Nick Bitzer and Alex Abbott



already compressed three hours in theprelims. As such, prioritizing what to testwas even more of a challenge. Before weeven knew exactly what the app was,going purely off the knowledge that itwas Android, the team had alreadydecided upon a few universal tools thatwe were going to use to do some staticanalysis on the .apk file, as well as othertools to do automated testing.

I grabbed the .apk file, tossed it intoFirebase (a platform developed byGoogle to help build quality applica-tions), and used the Firebase Test Labto run some automated tests against theapplication. I used the built-in automat-ed “robo” tests that come with Fire-base. I honestly wasn’t expecting toomuch to come out of the Firebase test,but since it took very little effort andtime on my end, I figured it couldn’thurt. In the end, it didn’t surprise methat Firebase didn’t turn up any bugs.

Another automated tool my teamused was the open-source MobSF(Mobile Security Framework). Wemainly turned to MobSF to uncoversecurity flaws in the application code.Unlike Firebase, MobSF actually turnedup some results, the best of which wasan SQL injection vulnerability—a bugthat would ultimately lead to us winningthe “Holy Cow Bug Award,” a distinc-tion for finding the most difficult bug.

We also leveraged Exer-ciser Monkey, a UI exercisertool within Android Studioto try to find bugs. In ourpreparation for the WorldCup, we used Monkey withgreat success on other dum-my apps, but unfortunately(or perhaps fortunately forthe product owners) had nosuch luck when using itagainst the Moovel app.

After the automatedsteps finished running, itwas “all hands on deck” forexploratory testing. The application hadsome well-defined functions, so it waspretty easy for the team to decide whowas going to test what once we had theapplication in our hands. For the nextcouple hours, we peeked into every cor-ner of the application trying to find asmany bugs as possible, although wethankfully didn’t have to repair any ofthe bugs we found.

We ran into a couple major obstaclesright away. Only two team membershad local cell phone service on theirAndroid phones, and part of the testingrequired us to leave the building and goout to taxicabs to make sure we couldreserve them. Because of this unexpect-ed snare, only half of the team was ableto test all portions of the application.

Individually, I also ran into an expired

trial period license forFirebase. I had been usingit to do some prep testing,but didn’t realize that I wasrunning a trial version of it,and I now had to pay forthe service. I (of course)didn’t have my wallet onme, so I had to track itdown and pony up for thefull Firebase subscription.Lesson learned: Alwayscheck the terms and condi-tions for your software.

I noticed the approach-es the competition used to write up bugs.Some would keep a running list of bugson paper and then write them all in onefell swoop toward the end of the exercise.Other teams, including ours, wrote bugsas they came up. We made a conscious,collective decision to handle it this way sothat the recreation steps were fresh inour minds and the logs for each failurewere readily available.

Who won?After all was said and done, we were ableto rack up a count of 39 bugs. Theyranged from suggestions on user experi-ence design to Java threads stuck in end-less loops. As a team, we felt very positiveabout the bugs we submitted and felt asthough our test report was comprehen-sive and laid out a clear overview ofeverything we thought should be fixed.

Frankly, I was just relieved it was allover—the weeks of staying up late, por-ing over practice application code, andthe stress and energy that were neces-sary elements of being ready to testanything and everything.

Although we ultimately didn’t end uptaking home the title of “Best SoftwareTesters in the World,” the entire compe-tition was an absolutely unforgettableexperience. We met a lot of amazing testengineers in the process, among themthe Pan Galactic Gargle Blasters, theteam from the Netherlands that earnedthe championship. They were fantastic,and we feel privileged to have competedalongside them. z

The Software Testing World Cup chronicles< continued from page 29

In the thick of searching for bugs in the Moovel app for Android.



http://www.textcontrol.com


BY CHRISTINA CARDONZA

CollabNet is joining the DevOps move-ment. The company announced a newDevOps solution and partnership at theDevOps Enterprise Summit.

“The No. 1 driver of DevOps in thebusiness is the desire of software develop-ers and the IT department to move at thespeed of business and to drive innovationfaster,” said Thomas Hooker, vice presi-dent and strategic market expert at Collab-Net. “We can no longer wait nine monthsto a year and a half to release new function-ality to our customers.”

The DevOps Lifecycle Manager isdesigned to help teams deliver high-qual-ity applications faster. It provides the abil-ity to see and trace across tool chains ofwork items, monitor events, and automateactions. In addition, it provides key per-formance indicators reports in order tohelp businesses identify and drive value.

“Increasingly the defining characteris-tics of so many products are software, andthe innovation speed required by the mar-ket is just getting quicker,” said Hooker.“DevOps mitigates risks by having smallerreleases and getting those releases outquicker and innovating quicker.”

The company also announced it isteaming up with DevOps softwareprovider Clarive Software in order to bringdevelopment and operations teams closer.“You have to be able to not just write code;

you have to deploy that code into produc-tion,” said Hooker. “We can write code fast.We can plan. We can test and we can deliv-er code fast, but...we can’t deploy into pro-duction. We have a bottleneck. We don’thave an automatic way to pick up code,take it, put it on the thousands of servers itneeds to go live on, and then if there is aproblem [of] how can I pull it back andmanage all those releases that are outthere.” The partnership with Clarive isdesigned to address that challenge.

According to Rodrigo Gonzalez, CEOof Clarive, the company focused on howoperations is handling DevOps today. Thepartnership will help the businesses retooltheir whole stack from ALM to applicationrelease automation, giving them full visibil-ity and generating insight to the business,developers and operators. The solutionfeatures demand management, monitor-ing, configuration management, deliverymanagement, traceability phase develop-ment, planning, rollback, and code quality.In addition, Clarive allows users who arestill doing waterfall approaches to benefitfrom application release automation.

“We are continuing to see DevOpsbecome not a fringe thing, but main-stream,” said Hooker. “One of the thingsthat makes us excited is how many existingenterprise and companies are looking toadopt DevOps and looking for help on howto be faster.” z

CollabNet enters DevOps arenaNew offering focuses on speed for building apps

DEVOPS WATCHDEVOPS WATCH

■ CA Technologies has announced

additions to its DevOps portfolio in

order to improve user experience.

The release includes updates to its

application performance manage-

ment and virtual network assurance

solutions. Features include real-time

insights, automated workflows, con-

tinuous testing, software-defined

networking support, and cloud con-

nectors. The company also recently

acquired Automic in order to

address IT and DevOps challenges.

■ IBM has announced new Bluemix

services aimed at making it easier

and faster to develop apps in the

cloud. The new services enable

developers to access and construct

tool chains using DevOps tools such

as GitHub and Slack. Services

include IBM Bluemix Continuous

Delivery, tool-chain templates, an

availability monitoring service, and

a partnership with Slack to bring

Watson to Slack developers.

■ SourceClear has unveiled new

features to help DevOps teams build

secure software. The company

announced features like issues

reports and suppressions; project

and subpath enablement; and proj-

ect-specific reports. In addition, the

company launched the Dependency

Visualizer to help teams see what

dependencies exist within a project

or library.

■ XebiaLabs has announced the 6.0

release of its DevOps solution. The

latest update to the XebiaLabs

DevOps platform is designed to help

organizations with diverse teams

manage and gain insight into com-

plex app releases. The main high-

lights include multi-functional

release folders, flexible release tags,

improved dashboards for compli-

ance and security, and enhanced

productivity for large-scale deploy-

ments. z

In other DevOps news…


The digital transformation is some-thing of a dream for the corporateworld. Whether executives have

grown jealous of fast-moving competi-tors’ products, or managers are sick ofbroken processes and hand-driven prod-ucts, the end goal of just about everybusiness in the world these days is toincrease software development budgets,and to create more products that aredriven by bits rather than wheels.

Take, for example, carmaker Volk-swagen. For it the future holds terrify-ing prospects that can only be met withsoftware-based solutions. One majoraspect of this future is the active collec-tion of user data.

When Tesla Motors began shippingits cars in 2008, it quickly becameapparent that those vehicles were col-lecting large amounts of data. TeslaMotors founder Elon Musk famouslygot into a scrap with then British TVshow “Top Gear,” citing the car’s datastores as a method of refuting theshow’s relative dissatisfaction with theTesla Roadster’s reliability.

Today, that data is not just being col-lected onboard, it’s being sent directlyto Tesla, allowing the company to ana-lyze its entire fleet’s behavior and short-comings. While this has transformedthe way Tesla handles its customer serv-ice, it’s also thrown other automakersinto a situation where this type of datacollection is now table stakes.

That’s fine to say on paper, but actu-ally collecting millions of points of dataevery day from millions of users in mil-lions of locations around the world is thesort of shoot-the-moon problem thatwould keep software developers andarchitects working briskly for a decade.For Volkswagen, the imperative is agreat deal closer to the windshield.

Roy Sauer is the man at Volkswagentasked with figuring out how to turn anautomaker into a software maker. As IT

CTO, he is responsible for the company’sinternal systems and its customer-facingapplications. Volkswagen has anotherCTO tasked with the technology of auto-mobiles, but Sauer is the one forced toface a growing influx of data, and anexpanding need for software.

Sauer said that Volkswagen createdits first customer-facing application thispast fall. The app helped Volkswagenowners work with their dealers to par-ticipate in the company’s massive recallof its diesel fleet in the United States.

While this was just a small app witha single purpose, Sauer’s teams are nowpivoting to meet demands placed upon

him by the business. One of thosedemands is to grow the company’s soft-ware revenue from zero to multiple bil-lions of dollars sometime in the nextdecade.

That’s a major requirement that willrequire Sauer and his teams to be thedrivers of the digital transformationinside of Volkswagen. Thus, workingwith Pivotal, Volkswagen has had todramatically redesign its softwareorganization. “As we are at a pointwhere we change completely, we nowwant to move forward to a much moreenhanced and increased softwaredevelopment approach,” he said. “Until


Companies can expect great rewards for cterrible consequences for falling short

Digital transformationto the future o

iBY ALEX HANDY

SDT332 page 34-36_Layout 1 12/20/16 3:52 PM Page 34


now, we had much less capacity in soft-ware development.”

Given success on the app front, onlyhalf the battle is won: Sauer still has todeal with the influx of all that driverdata over the next four years. Managingthat data is a big part of his road map.

“This is one of the hardest topics,” hesaid. “What we know is that we will havea huge volume of data, and that is a bigasset. It will be very important that wecan handle this data. It starts again withour cars: We will produce as a group 10million cars per year, and from 2020 onwe will have 10 million connected carsper year, and we will have a huge num-

ber of devices in the Internet, like sen-sors for autonomous driving. They willproduce a huge volume of data.

“The challenge is how to handle thisdata, how to use this data, and how toenhance further services for our cus-tomers. One critical challenging topicwill be who actually owns the data andwhat will be legal. Where do we have tostore this data? What is the data protec-tion policy in U.S., in the EU and inChina? This will be a really sophisticat-ed kind of data management we have todo there.”

Dam transformationOne successful digital transformationtook place inside BC Hydro, the powergeneration company in British Colum-bia, Canada. It decided to roll out IoTdevices to its entire power grid as earlyas 2005, but the real effort to do so didnot begin until 2011.

Pushing sensors into its network wasno small task. While the software devel-opment behind those sensors took plen-ty of time, the project had help from BitStew Systems, a consulting firm that wasacquired by GE. The efforts made oncomputer screens, however, wereinsignificant compared to the amount ofwork that was required in the field todeploy those sensors.

As British Columbia is quite largeand includes some rather remote areas,one sensor, for example, requires a fullday’s drive into the wilderness coupledwith a full day’s hike to it.

David DeYagher, manager at BCHydro, said that the entire projectrequired solutions that would last at least20 years. “Our program was going to costour company CA$900 million,” he said.

“With that we had to lay claim to$1.6 billion of benefits. The first partwas the reduction in system losses usingadvanced equipment, and to put outintelligent apps to support system loss-es. This would do things like help usfind where we had folks borrowingpower from us and not paying.”

Just as with Volkswagen, however,once BC Hydro had deployed 2 millionsensors to its grid, the real problemscame around the management andanalysis of the data.

“We went from 30,000 data points aday to over 270 million data points a day.We needed someone with experience interms of not only being able to ingest, butto also integrate some of the data into oursource systems to drive the value. Wealso realized we had a plethora of systemsour operators had to use. What the BitStew team offered was a single pane ofglass, a manager of managers for tier twoand tier-three operators have single paneof glass to understand every meteringdevice, every operating device that drovea lot of value for us.”

Which brings the digital transforma-tion story back to the data, where manyhave said that the transition shouldbegin. Jack Norris, CMO of MapR, saidthat bringing the data together can bringthe organization as a whole together.

“The reason for convergence is thatthe digital transformation is about con-verging operations and analytics togeth-er: Combining analytics with operationsso you’re not reporting; you’re adjustingthe business as it happens,” he said.

“Less attention has been focused onthe underlying data layer, but it’s theunderlying data layer that enables real-time and missions critical capabilities.If you look at the dimensions of speed,scale and reliability, it’s very difficult todo all three across a large distributedcluster. I think regardless of what youcall it,” said Norris, building a data lakecan draw developers, analysts and otherstake holders in to a single place to gettheir data.

“That’s just a small step on the jour-ney, and by no means the stoppingpoint because it’s about the ability toleverage that data in context and in realtime. It’s the enabling layer for thesetypes of apps and transformations.That’s where the second key of streamprocessing comes in,” said Norris.

“It’s typically these event-based dataflows that are important to harness,whether it’s web events and machinesensors, or biometrics or mobile. If youcan leverage those events and leveragethose in context, and provide the histor-ical aspects of data in motion as well asdata at rest, that solves the problem offailure alert, monitoring and adjusting,


r completing it, and

e of businessis essential



supply chain optimization, ad optimiza-tion—the list goes on and on.”

That data is the heart of the transfor-mation is echoed by Shaun Connolly,chief strategy officer of Hortonworks.“The journey to digital transformationfrom our perspective is around the dataand how do you assemble it in a way tofuel the digital transformation initia-tives,” he said.

“For their digital transformation ini-tiatives, enterprises are focused oncloud, IoT and Big Data all swirlaround with the commonality beingdata. It isn’t necessarily about gettingdata in one place; it’s about getting datainto places it needs to be.”

Another enabler for the digital trans-formation, said Connolly, is the merg-ing of public and private clouds. Offer-ing internal developers a quick and easyway to provision their systems is impor-tant, but being able to scale those sys-tems into public clouds when needed isjust as crucial, he said.

“That connected architecture onpremise and in cloud...we’re increas-ingly seeing this as a digital transforma-tion architecture,” he said. “You getsome analytics, and in many cases wantto act on real-time data and assemble itfor cloud-based analytics and machinelearning scenarios for a 360 view onroot causes, in a manufacturing orrepair type scenarios.

“We’re seeing it when you look at thevideos of Hadoop Summit. We had wellover 20 different customers sharingarchitectures, like Comcast, Ford orCapital One. When you look at their dig-ital transformation initiatives, they’reincreasingly inherently hybrid. It isn’thybrid around ‘Let me burst my data tothe cloud.’ It’s ‘How do I get my data setswhere they need to be in a way that’ssecure and well governed?’ Data sciencecan explore, but you can also do that 360-degree view of customer on premisebecause you can mix it with data from theenterprise that may or may not makesense to put in the cloud.”

Customer transformationAdam Seligman, executive vice presi-dent and general manager of Salesforce

App Cloud, said that while digital trans-formation is a murky prospect, manycompanies are diving in headfirst.

“I'm actually blown away by howcompanies are leading into this cus-tomer-centric transformation,” he said.“It’s happening in every industry, andevery segment. Banks and healthcareservice companies are all reinventingthemselves in the age of the customerwith the concept of digital transforma-tion. I had a financial services institu-tion said they used to wrapped theircustomers around their products, butnow they want to wrap their productsaround their customers.”

The demand for the digital transfor-mation of enterprises is not somethingthat has been creeping up on companiesfor years, either, said Seligman. It’s some-thing they suddenly realize is a businessnecessity: One day a competitor makes ita priority, or perhaps there’s a top-downmandate. However it happens, it seemsto be more of a race than a saunter.

“They wake up one day and suddenly

they're in the software business. Theseare 150 year old companies and theyhave to rethink their business with anapp at the center of it. Their productsand services are all threaded throughthis. It’s not just one app, it’s a wholecloud of apps. It’s connecting the cus-tomer with the community they partici-pate in and making sure sales engages ina really customized way,” said Seligman.

“Fitbit is a brilliant device you wearon your wrist, but the true greatness ofFitbit is this community, the apps youuse, and the communications from Fit-bit keeping you motivated. That’s thebrilliance that keeps you using the Fit-bit,” said Seligman.

And that’s the overall goal of the dig-ital transformation: to turn great prod-ucts into great com-munities that react tothe customers’ needsin real time, withoutguessing and withouthumans driving everystep in the process. z

Read this story onsdtimes.com

We sat down with Colin Parris, Ph.D., vice president of GE Software Research at thecompany’s Minds and Machines conference to find out just how GE has navigated itsdramatic digital transformation, and to get some tips on how enterprises can movealong their transition path.

“How I think about digital transformation coming together is,I look at my processes. How do I make the process of designingmy gas turbines more effective? What we’ve done is built advisorsusing AI in software. When I actually try to design somethingwith the advisor, it tends to see things. What if the problem youhave with this design is it’s not manufacturable?

We also use the AI to do a lot more simulation than before. I would run 50 combustion experiments before because to firethis up at a big site takes a week to set up the experiment. NowI am using AI when we consider doing 50: Here’s the next eightyou can do. You can simulate the points between and you can test in the middle.

In the data, I can realize two things. One is that when I would bring in the jet enginesevery 500 flights to check, I can do a digital twin based on the data I collect. What do Irepair? That’s the digital transformation: transforming the key processes by usingknowledge I have and by using machine learning and AI techniques to predict the future.

Most people collect data, but it’s never tagged. I collect from five sources, but whenI ask you to show me the sources, we find one sensor is dead, or part of the data is cor-rupted, or the timestamps are off. I need to know if this data came out of a good system.

You’ve got to make sure you’ve normalized and cleaned the data. Collect it andstore it, but it could be useless because someone flipped a switch on a machine. Takesome time...to run some initial analytics to see how useful the data is.

Finally, try to figure out what’s the value in the data... It may be very valuable toknow the humidity, but if you didn’t save the humidity, the entire value is lost. Take itwith the value in mind.” z —Alex Handy

IoT as digital transformation enabler

Colin Parris




http://www.progress.com/futureproof-web-apps

SDT333 Full Page Ads_Layout 1 12/22/16 10:17 AM Page 38

CLICK

CLICK

CLICK

CLICK

http://resources.sdtimes.com/newsonmonday

http://resources.sdtimes.com/andevnews

http://resources.sdtimes.com/interdronenews

http://resources.sdtimes.com/sptechreport

http://www.sdtimes.com

INDUSTRY SPOTLIGHT: API MANAGEMENTINDUSTRY SPOTLIGHT: API MANAGEMENT


BY LISA MORGAN

APIs are the building blocks of today’sdigital economy. Although more organ-izations are building APIs with the goalof driving more value from their digitalassets, many of those companies havetrouble managing their APIs effectively,especially at scale. Red Hat brings orderto API chaos so software teams canspend more time creating tangible busi-ness value.

“You can’t just create an API andthink you’re done with it,” said SameerParulkar, product marketing manager,Enterprise Middleware, at Red Hat,Inc. “You may create an API for a par-ticular purpose today, but what aboutwhen requirements change tomorrow?How will you manage and secure thatAPI? You need a scalable, enterprise-class way of doing all that.”

Managing and securing APIsrequires a platform capable of provid-ing caching, fault tolerance, trafficrouting and load balancing. Red Hat3scale API Management Platformscales to billions of calls and its distrib-uted architecture ensures no singlepoint of failure.

3scale also provides enterprise-classAPI access control and security. With it,companies are able to provide differentlevels of access to different types ofusers. They are also able to control howdifferent applications interact with aparticular API. 3scale’s powerful APIaccess, policy and traffic controls makeit easy to authenticate traffic, restrict bypolicy, protect backend services, imposerate limits and create access tiers. It alsohas the analytics developers need tomonitor trends and peak usage times.The analytics also help developers bet-ter understand which applications gen-erate the most traffic, which APIs aremost popular, and which APIs or end-points are used the least.

Businesses that want to create newrevenue streams from their APIs can use

3scale’s platform to set pricing rules,invoice customers, and collect payments.

Enable Real-Time IntegrationAn enterprise’s information is spreadacross multiple applications and process-es, which makes APIs complex and diffi-cult to manage. To address the problem,organizations need an effective means ofintegrating applications, data, anddevices within and beyond the enter-prise. Progressive companies chooseRed Hat JBoss Fuse because it enablesrapid integration across extended enter-prises on-premises or in the cloud.

“Today’s APIs use multiple informa-tion sources, so organizations need away to connect those informationsources together in a modular way,”said Parulkar. “Red Hat JBoss Fuse is arobust, flexible, and easy-to-use plat-form for integrating applications, dataservices and devices. It uses popularopen source technologies like ApacheCamel to provide transformation, rout-ing, and protocol-matching service.”

Red Hat JBoss Fuse also includesRed Hat JBoss A-MQ, a high-perfor-mance messaging platform based onApache ActiveMQ. Its integration capa-bilities eliminate manual touch points,automate processes, and connect enter-prise assets for improved efficiency. Italso includes the real-time messaging

capabilities agile businesses require.As companies move to microservices

architectures, they need the ability tocreate and connect an API in an agile,continuous fashion. To achieve that, theyneed a platform that supports the cre-ation of the API as well as its deploy-ment and redeployment. Red HatOpenShift Container Platform is a com-prehensive application development andhosting platform that automates tediousmanagement tasks so developers canspend more time building apps thatalign with business goals.

Red Hat OpenShift also facilitatesDevOps by enabling development andoperations to work faster and more effi-ciently. Like all Red Hat products,OpenShift provides enterprise-gradesecurity and improves the effectivenessof DevOps. The platform enablesdevelopment teams to develop, deploy,and automate modular, reliable, andserviceable applications across theirinfrastructure regardless of the applica-tion’s architecture.

Notably, OpenShift integrates all ofthe architecture, processes, platformsand services needed to empower devel-opment and operations teams. Finally,businesses can create ideal cloud com-puting services tailored to applications,integration, or mobile.

Mobile, social computing, and thecloud require organizations to inte-grate flexible software into the veryfabric of what they do. To succeed, theAPIs they build must be able to accesscore systems and resources that mayhave become information silos. Byexposing data, business processes, serv-ices and resources through APIs, com-panies are in a better position toimprove mobile experiences, growtheir ecosystem, expand their reach,power new business models and cat-alyze internal innovation. z

> LEARN MORE about Winning in the API Economy.

Red Hat powers the API economy

Content provided by SD Times and

Sameer Parulkar

SDT333 page 39_Layout 1 12/27/16 9:56 AM Page 39

http://pages.3scale.net/rs/516-GHI-083/images/Winning-in-the-API-Economy-eBook-3scale.pdf

Find out how Sauce Labscan accelerate your testingto the speed of awesome.

For a demo, please visit saucelabs.com/demoEmail [email protected] or call (855) 677-0011 to learn more.

B E F O R E S A U C E L A B SDevices. Delays. Despair.

A F T E R S A U C E L A B SAutomated. Accelerated. Awesome.

A brief history of web and mobile app testing.


http://www.saucelabs.com/demo

Companies today are liketeenagers learning how to drive;there’s a lot of stop and go, and

a strong desire to go fast. While it’s nec-essary for organizations to find a solu-tion that fits their agile transformation,experts suggest there is no need to rush.Before getting the green light to go tomarket, teams should slow down andtake a step back so they can avoid somemajor automation pitfalls.

Like anything with software testing,there is risk. Automating tests takeseffort in order to be successful, and thelast thing business leaders want arepoorly conceived and executed tests,creating a costly and time-wasting mess.

Lubos Parobek, vice president ofproduct at Sauce Labs, said a major riskhe sees companies concerned with is theidea of test flakiness, where tests failbecause they were badly written. This iswhy best practices are important, he said,because the more teams are educated,the better chance they have at buildingtests that are maintainable and reliable.

Michael Eckhoff, software testautomation veteran at Tricentis, said oneof the biggest pitfalls he sees is compa-nies trying to pick a software automationtool and expecting it to solve all of their

problems. What companies need to do isfigure out how to pick the tool that bestmeets their needs, and to ensure that thecompany is ready for it, he said.

“You can typically get that first quickwin with intuitive test-case design,” saidEckhoff. “Getting to that second phaseof really intelligently designing your testportfolio to ensure that you are testing toget the best risk coverage or best busi-ness value is where that [best-practice]methodology comes back into play.”

Another risk comes from overconfi-dence, said Eckhoff. Companies look attheir processes, see a large amount ofautomated test cases, and assume theyare in “good shape,” he said. The prob-lem in this case is the tests are not clear-ly defined, and in many cases, the testdefinition doesn’t stay consistent withwhat the test is actually doing, he said.

Companies need to get their teams tobuy into the concept of test automation,or organizations will continue to treatagile as the “Wild West,” where teams dowhatever they want just to release thingsfast, and then when it breaks, they do itagain, said Eckhoff. He said this way ofthinking just doesn’t work for the enter-prise, and those organizations needmore than just a tool to achieve speedand quality applications; they need aprocess in place that supports agile

development and QA early on.

Rushing to automationWhen getting started with a softwareautomation tool or solution, organiza-tions should pause and do a bit moreplanning up front before jumping intothe deep end, according to Parasoft’ssoftware industry expert, Marc Brown.He said he has seen organizations selectan automation tool based on a very lim-ited developer group, but in the end,the assessment wasn't broad enough tosee how it impacted their daily activitiesor how it helped their agile initiatives.

Brown also said that a lot of teamsget into problems with their testingstrategies because they rush into usingan automation tool to quickly solve oneproblem, which means they lack the“big picture,” or what software automa-tion can do over time for their company.

“I think if operations slows downand they get guidance from people whohave expertise in agile development[and] Continuous Testing, and havelooked [at the tool] broadly to see howit supports the developers, the testers,the managers, and the executives,” saidBrown, “most people will recognizethere is value in getting a solution thatcovers all of those various roles.”

Move fast while avoidingautomated testing

pitfalls


BY MADISON MOORE


Buyers Guide

Slowing down and cutting corners are not options, and neither is

overconfidence about your processes

SDT333 page 41,42,45,46_Layout 1 12/20/16 3:54 PM Page 41

mara

Highlight

mara

Highlight

mara

Highlight

http://www.saucelabs.com

http://www.tricentis.com

http://www.parasoft.com

Brown added that the one shortcuthe sees companies taking is not testingadequately. He said these companies“race to market with blinders and reallycross their fingers hoping they won’t getcaught up with a bad defect.” The riskof not testing can be as detrimental ashurting the company’s brand or marketreputation. And in some cases, having amalfunction in software could really

physically hurt the customer. Parobek said that he also sees com-

panies trying to cut corners, and in hispersonal experience, it’s never worth it.

“Consumers more and more havevery high expectations of web and mobileapps that are coming from the servicesthey use, whether that’s banks, retailers,etc., so having the application that youcut corners with that you deliver earlyand it is buggy, it’s a bad tradeoff,” said

Parobek. “Going to Continuous Delivery,Continuous Integration, and [automat-ed] testing is a way you can speed every-thing up and get higher quality.”

He does believe that with the righttool, companies can“have your cake andeat it too,” but thatrequires proper invest-ment and planning todo it right. z

Marc Brown, software in -

dustry expert, Parasoft

Parasoft helps organiza-tions perfect today’s highlyconnected applications by

automating time-consuming testingtasks while providing management theanalytics necessary to focus on whatmatters.

We also help developers becomehighly efficient at software validationusing various technologies and tech-niques—from static and runtime analy-sis, to unit testing, to API testing all ona scalable software virtualization plat-form. Furthermore, we help the man-agement team focus on what mattersthrough highly unique data analyticsand reports. By doing so, the teameliminates issues that will impact qual-ity, security, performance, reliability,and customer satisfaction.

Parasoft solutions help teams elimi-nate defects early in development min-imizing cost and rework, and focus onthe risky areas of an application, savingtime while improving quality and low-ering security threats.

Lubos Parobek, vice

president of product,

Sauce Labs

The right tools are criticalfor a successful outcome

when embarking to adopt continuousdelivery and automated testing. Forexample, many organizations strugglewith getting the required test coverageacross platforms, while keeping testingfrom becoming a bottleneck in theirdelivery pipeline.

Many also struggle with higher errorrates and flaky tests, particularly when

the flakiness is caused by a failure intheir in-house test infrastructure, notthe test code itself. Sauce Labs can dra-matically increase the speed, coverageand accuracy of automated testing. Forexample, Sauce Labs allows our cus-tomers to run tests at much higher lev-els of concurrency, with better coverageand with lower error rates than mostcan achieve on their own. We run over1.3 million tests a day across 800+ OS,Browser and Device combinations.

Gerd Weishaar, vice

president of product

management, Tricentis

It’s become increasingly dif-ficult for teams to build a

test suite that not only covers the risksthat are most critical to the business, butalso is maintainable and efficient enoughto be used for continuous testing of rap-idly-evolving applications. Deliveringpositive user experiences requires con-tinuous testing—and accelerated devel-opment cycles mean there’s less timeavailable for designing, creating, updat-ing, and executing tests.

Our solution is designed to helptesters rapidly identify and create thetests that are most important for pro-tecting key end-to-end businessprocesses. Based on many years of R &D focused on Model-based TestAutomation, we’re able to scan theapplication under test to help testersunderstand and define the optimal setof test cases that will cover their orga-nization’s highest priority risks. Theresulting test suite is perfectly-suitedfor continuous execution as part of thedelivery pipeline, and it’s easily updat-ed as the application evolves.

All this is accomplished withoutcoding or scripting. Scripted tools leadto long complex test scripts, which get“fixed” in the script if technical chal-lenges arise. At some point thesescripts become too many, too complexand end up as maintenance nightmare.This is typically the reason why manualtesting is still so popular (80% of testcases) after 20 years of (script based)automated test tools like QTP/UFT.

Rod Cope, CTO of

Rogue Wave

Developers and QA engi-neers have to considerwhat might go wrong with

code, but they aren’t able to anticipateeverything. Despite best efforts, they’rebuilding in and shipping bugs and secu-rity vulnerabilities with code.

Klocwork, our automated staticanalysis tool, enables software teams tofind and fix bugs early in the life cycle,which saves time and money whilespeeding software delivery. It continu-ously monitors every line of code soerrors can be identified that developersand QA engineers miss. Klocwork canbe integrated seamlessly with continu-ous integration environments so issuescan be identified and remediated priorto a build. As you’re writing code, it willtell you if you’ve introduced an error soyou can fix it immediately. Klocworkalso improves workflows and enforcesbest practices by enabling softwareteams address issues earlier in the lifecycle.

Klocwork users are able to preventsecurity exploits and compliance issuesthat can negatively impact brands, cus-tomer relationships and profitability. z



What testing pitfall does your solution aim to solve?



The new Gartner Magic Quadrant for Software Test Automation 2016 is here!

Gartner recognizes Tricentis as a Continuous Testing Leader

Download your copy now!www.tricentis.com/GartnerMQ2016

Tricentis


http://www.tricentis.com/GartnerMQ2016

SDT333 Full Page Ads_Layout 1 12/21/16 11:09 AM Page 44

http://www.parasoft.com/forrester-wave


n Applause delivers unmatched in-the-wildtesting, user feedback and research solutionsby utilizing its DX platform to manage commu-nities around the world. The company’s testingsolutions span the entire app lifecycle andinclude access to its global community ofmore than 250,000 professional testers inmore than 200 countries and territories aswell as access to specific, client-requested dig-ital users.

n CA BlazeMeter: The solution ensuresfaster delivery of applications by enablingDevOps teams to quickly and easily run per-formance tests on-premises, in the cloud oron virtual private clouds against any app,website or API at massive scale to validateperformance at every stage of softwaredelivery.

n HPE Software’s automated testing solu-tions simplify software testing within fast-moving agile teams and for Continuous Inte-gration scenarios. Integrated with DevOpstools and ALM solutions, HPE automated test-ing solutions keep quality at the center oftoday’s modern applications and hybrid infra-structures.

n LogiGear: With the no-coding and key-word-driven approach to test authoring in itsTestArchitect products, users can rapidly cre-ate, maintain, reuse and share a large scale ofautomated tests for desktop, mobile and webapplications.

n Microsoft provides a specialized toolset for testers that delivers an integratedexperience starting from agile planning totest and release management, on premis-es or in the cloud.

n Mobile Labs’ deviceConnect providessecure remote access to mobile devices formanaged use by developers and testers. WithdeviceBridge (an extension to deviceCon-nect), many test automation frameworks anddeveloper tools used for checkout and debug-ging can retain cloud-based devices as iflocally connected by USB.

n Neotys load testing (NeoLoad) and per-formance monitoring (NeoSense) productsenable teams to produce faster applications,deliver new features and enhancements inless time and simplify interactions acrossDev, QA, Ops and business stakeholders.

n Orasi is a leading provider of software test-ing services, utilizing test management, testautomation, enterprise testing, Continuous

Delivery, monitoring, and mobile testing tech-nology.

n QASymphony’s qTest is a Test Case Man-agement solution that integrates with popu-lar development tools. QASymphony offersqTest eXplorer for teams doing exploratorytesting. qTest Scenario is a JIRA add-on thathelps teams optimize and scale Test Firstmethodologies across their organization.

n SOASTA’s Digital Performance Manage-ment (DPM) Platform enables measurement,testing and improvement of digital perform-ance. It includes five technologies: mPulsereal user monitoring (RUM); the CloudTestplatform for continuous load testing; TouchT-est mobile functional test automation; DigitalOperation Center (DOC) for a unified view ofcontextual intelligence accessible from anydevice; and Data Science Workbench, simpli-

fying analysis of current and historical weband mobile user performance data.

n Synopsys: Through its Software Integrityplatform, Synopsys provides a comprehensivesuite of testing solutions for rapidly findingand fixing critical security vulnerabilities, qual-ity defects, and compliance issues throughoutthe SDLC.

n TechExcel: DevTest is a sophisticatedquality-management solution used by devel-opment and QA teams of all sizes to manageevery aspect of their testing processes.

n Progress: Telerik Test Studio is a test-automation solution that helps teams bemore efficient in functional, performance andload testing, improving test coverage andreducing the number of bugs that slip intoproduction. z

A guide to automated testing toolsn Parasoft researches and develops software solutions that help organizations

deliver defect-free software efficiently. To combat the risk of software failure while accelerat-

ing the SDLC, Parasoft offers a broad test. Parasoft's enterprise, mobile, IoT, and embedded

development solutions are the industry's most comprehensive—including static analysis, unit

testing, requirements traceability, coverage analysis, API testing, dev/test environment man-

agement, service virtualization and more. The majority of Fortune 500 companies rely on

Parasoft in order to produce top-quality software consistently and efficiently as they pursue

agile, lean, DevOps, compliance, and safety-critical development initiatives.

n Rogue Wave is the largest independent provider of cross-platform soft-

ware development tools and embedded components in the world. Rogue Wave Software’s

Klocwork boosts software security and creates more reliable software. With Klocwork,

analyze static code on-the-fly, simplify peer code reviews, and extend the life of complex

software. Thousands of customers, including the biggest brands in the automotive,

mobile device, consumer electronics, medical technologies, telecom, military and aero-

space sectors, make Klocwork part of their software development process.

n Sauce Labs provides the world’s largest cloud-based platform for auto-

mated testing of web and mobile applications. Its award-winning service eliminates the

time and expense of maintaining an in-house testing infrastructure, freeing development

teams of any size to innovate and release better software, faster. Optimized for use in CI

and CD environments, and built with an emphasis on security, reliability and scalability,

users can run tests written in any language or framework using Selenium or Appium, both

widely adopted open-source standards for automating browser and mobile application

functionality. Videos, screenshots, and HTML logs help pinpoint issues faster, while Sauce

Connect allows users to securely test apps behind their firewall.

n Tricentis is recognized by both Forrester and Gartner as a leader in software

test automation, functional testing, and continuous testing. By enabling test automation

rates of over 90%, we help Global 2000 companies control business risk as they adopt

DevOps, Agile, and Continuous Delivery. Our integrated software testing solution, Tricentis

Tosca, provides a unique Model-based Test Automation and Test Case Design approach to

functional test automation—encompassing risk-based testing, test data management and

provisioning, service virtualization, API testing and more. Tricentis’ 400+ customers

include global names from the Top 500 brands such as A&E, Allianz, BMW, Deutsche Bank,

HBO, Lexmark, Orange, Starbucks, Toyota, UBS, Vantiv, Vodafone, and Zurich Insurance.

Rogue Wave

Parasoft

Sauce Labs

Tricentis

n FEATURED PROVIDERS n


mara

Underline

mara

Underline

mara

Underline

mara

Underline

http://www.saucelabs.com/demo

http://www.tricentis.com/GartnerMQ2016

http://www.parasoft.com

http://www.klocwork.com/


BY LISA MORGAN

When developers and QA engineerstest code, they have to imagine whatmay go wrong in production. Even ifthey do a code review or read throughthe code, there are always bugs that arevery difficult to detect. To ensure thosebugs can be identified and remediatedquickly, software teams are embracingRogue Wave’s Klocwork, an automatedstatic code-analysis tool that ensuresevery line of code, every function ormethod call, and every parameter hasbeen checked so errors can be fixedbefore a build completes.

“Developers and QA people don’thave the time and focus to test every sin-gle line of code in an application. WithKlocwork, you have an automated way todo it so you can ensure that nothing slipsthrough the cracks,” said Rod Cope,CTO of Rogue Wave Software. “It’simportant that quality and securityissues don’t sneak in because developersare working at 2:00 a.m. to meet a dead-line. Klocwork always has your back.”

Deliver higher-quality code fasterAs teams add more features to theircode, deliver it faster, and adhere tostricter standards, finding and fixingsecurity flaws becomes more difficult.The vulnerabilities can result in databreaches and application crashes whichcould have been prevented early in thesoftware life cycle.

Klocwork identifies many hard-to-find issues in code such as buffer over-flows or buffer overruns, memory leaks,deadlocks, multithreading code issues,and compliance issues. Using Klocwork,developers can avoid security exploitssuch as Heartbleed while improvingsoftware quality and the economics of

software delivery.For example, Lawrence Livermore

Labs saved $200,000 on a small project,and Harris, a defense contractor and anIT services provider, saved $60,000 on apilot project. One customer identified20% more bugs in its IoT code. Anotherexperienced a 90% increase in the linesof code per developer when integratingKlocwork with its Continuous Integra-tion solution and running the analysis.

“Klocwork almost doubles thecapacity of your team. If you add all theefficiency gains up, it’s like doublingyour engineering team without hiringpeople,” said Cope.

A recent Rogue Wave survey foundthat most developers are responsiblefor securing the software they producebecause their companies lack theappropriate security personnel. Anoth-er survey found that 80% of developersresponding admitted they don’t knowhow to secure software.

“Everybody needs help with securitybecause it’s such a big problem,” saidCope. “We can automate that so ourcustomers can be sure that their appsare properly secured regardless ofwhether or junior developer or seniorperson is writing the code.”

Security isn’t the only challenge,however. Developers are doing moretypes of testing than ever while softwaredelivery cycles continue to accelerate. Iferrors can be identified prior to a build,less testing will be required later, reme-diation costs will decrease dramatically,and software can be delivered faster.

“Today’s development teams areonly writing 10% to 20% of the code intheir app, and the rest comes fromopen source, contractors, offshoring,and on-shoring. Other parties are writ-

ing code that’s used in the apps,” saidCope. “Static code analysis can test allthat code—not just the code your teamhas control over.”

Build quality into Continuous IntegrationThe best way to build quality into Con-tinuous Integration processes is toensure seamless integration of the test-ing tool and the Continuous Integrationsolution. Klocwork provides that inte-gration.

Some Rogue Wave customers havemade static code analysis part of thebuild and test flow, so a build will fail ifa security defect is identified during thecode analysis scan. When such an issuehas been identified, Klocwork notifiesthe developer who created it so to theissue can be remediated before thecode is checked into the main line.

Klocwork also notifies developers oferrors they’re introducing as they writecode so they can be fixed immediately.For example, if a developer types in anerroneous syntax that introduces asecurity vulnerability, a red squigglewill appear below the offending code.That way, the developer can correct theerror in context rather than waiting forthe code to go through a QA cycle, get-ting familiar with the code again, andthen resolving the issue.

“Klocwork makes static code analysishassle-free for developers. It enablesthem to prevent errors upfront ratherthan fixing them later,” said Cope. “Foryears, developers have been told thatfixing bugs earlier in the life cycle savestime and money. Not everyone knowsthat static code analysis takes the guess-work out of testing so higher-qualitysoftware can be delivered faster.”

Learn more at roguewave.com. z

Rogue Wave ensures quality, security

Content provided by SD Times and



http://www.klocwork.com/sca

48 SD Times January 2017 www.sdtimes.com

Guest ViewBY ALAN HO

When a company embarks on “digital transfor-mation,” it often has to modify its software

systems. This can become excruciatingly difficult inlarge organizations with monoliths—large, custom-built software systems with multiple developmentteams working on the same codebase.

A common response from companies with theirbacks against the “monolithic wall” is to hire execu-tives from well-known tech companies like Google,Microsoft, Amazon and others to “digitally trans-form” them. This often involves moving to a newarchitecture that leverages microservices and APIs.

Frequently, when this new technical leadershipcomes in, they bring their most trusted developerswith them. This can help speed up the transition tomicroservices, but there’s also a shelf life to how longgood developers are willing to stick it out. If themonolithic stack doesn’t go, the good developers will.

Trouble is, monoliths are pesky,and replatforming isn’t always aseasy as it seems. At the end of theday, the transition to microserviceshas less to do with technology andmore to do with managing peopleand vendors. Leaders who thinkthey can build new software and

avoid touching legacy systems and confronting orga-nizational roadblocks are fooling themselves. There’ssimply too much inertia to avoid these challengesand replatform to a new architecture the right way.

While failure to make the transition can amountto career suicide, recognizing the real challenges canlead to success.

Embracing vs. resisting change: Like it or not,there will always be two groups of developers in yourorganization: Those who drive change, and thosewho resist it at every turn. Competent and well-intentioned as the reisters may be, getting them tofollow the new microservices model is a cultural shiftthat must happen—and sooner rather than later.

The best way is to overcome this is “leading byexample” by having those early adopters roll out afew microservices and show that they are indeedeasier to operate and develop than monoliths. Areasthat may immediately chafe existing developers andoperations members are concepts such as “develop-ers doing operations,” or that each microservicesteam is responsible for both the “operational” and

“business” success of their own service. Taking control of your software stack:

Entrenched software vendors of web applicationservers, ESB and virtualization technology oftendrop roadblocks in front of technical leaders whenthe only choices for building microservices areopen-source and cloud-native technologies. Thesevendors may pair up with “change resisters” withinan organization to block change.

To avoid this, leaders should give their earlyadopters carte blanche control of the stack whilenegotiating down the cost of their entrenched soft-ware vendors. Negotiating down costs, or eliminatingthem entirely, can be used to fund the “new” soft-ware stack or other efforts such as automation.

Cloud migration imperative: Microservicesrequire a significant increase in capital and operatingexpenditures because each microservice needs itsown cluster of servers to run on. The move frommonoliths to microservices means the number ofclusters required to run a microservices architecturemight increase the number of servers by 10x, or even100x. Without moving to the public cloud, or trans-forming the data center to run cloud-native technol-ogy like Cloud Foundry, the CAPEX/OPEX costscan be totally prohibitive.

With this in mind, IT leaders should set expec-tations that upfront investment is necessary. Insome cases, this investment is spent on setting up apublic cloud instance. In others, on transformingexisting data centers with platforms like CloudFoundry. The second part of investment is oninfrastructure automation so that it becomes costeffective to manage all these microservices at scale.

Make fear your ally: When it comes to migra-tions, fear sometimes works better than encourage-ment. To kill the monolith at Amazon, an edictcame down saying that whichever team failed todecouple themselves from the monolith would endup owning and supporting the monolith. No devel-oper in their right mind wants to own a bunch ofcode that was written by someone else, so the rightincentives were put in place for the transition.

Taking into account the organizational challengesof moving to microservices is just as important as thetechnical challenges. But with just a bit of upfrontthinking and planning, most enterprises can navigatethis transition with a lot less pain and angst. z

Diffusing the monolith time bombAlan Ho leads developerrelations at Apigee.

The organizational challenges

of moving to microservices

are just as important as the

technical challenges.




Analyst ViewBY AL HILWA

The 2016 election took place while I wasattending a tech conference in Seattle. Look-

ing around and watching my messages and socialmedia, it was clear that many in the tech commu-nity were deeply disturbed by the result.

I wanted to take a moment to share my perspec-tive as an American who immigrated to this coun-try a few decades ago. I came to this country as acollege student in the early 1980s. I knew littleabout this country and even less about its politicalsystem. I did my studies and worked through mydegrees in the heartland of America (Wisconsin,then Missouri), gaining a deep appreciation ofwhat makes this country strong.

By the definition of left/right politics, I was apo-litical, though I noticed the sentiment againstRonald Reagan in academic circles and confess todeeply enjoying the movie “Being There,” featur-ing the great Peter Sellers allegorically lampooningwhat was assumed to be Reagan’s serendipitousrise to the presidency. Many years later I noticedthat Reagan was increasingly being recognized astransformative for the country in multiple ways.

While there are differences in experience andpersonality between Reagan and our current presi-dent-elect, the moral of the story is to expect theunexpected and to understand the huge biases thatlurk in your outlook and that of your tribes. Leadershave been known to grow and evolve on the job.

One thing that I learned is that most people inthis country are quite apolitical; they move throughtheir lives with rightful disdain for what goes on inWashington as they produce innovation and eco-nomic value. The wisdom embedded in our systemof governance has proven resilient and fosterspeaceful transitions between rulers, allowing therest of us to focus on what really counts in keepingthe country strong. I am not advocating political dis-engagement, but reflecting on whether we haveallowed ourselves to be too politically vested is due.

America, the innovation magnetOver the years, I have learned that the diversity ofthought in this country is its greatest source ofstrength. This diversity is fed and nurtured by a his-tory of immigration of diverse groups and supportedby a natural variety in the country’s geography. Thisdiversity, when combined with a mild regulatory

environment and the natural wealth of the countrythat has led to strong investment in institutions andbusinesses, has accounted for its economic successand business leadership in technology. In addition todiversity of thought, immigration will be crucialbecause native fertility always declines with personalwealth and industrialization.

It is interesting to observe the conversationaround the inevit able rise of China toppling theU.S. in its economic standing. While I have mydoubts on this and the linear thinking behind it, Ihave to note that the key edge that China has overthe U.S. is population—plenty of it. The U.S. willstand a better chance of keeping up with China ifit continues to embrace a vigorous immigrationstance. My expectation is that the new presidentand his people understand this and that our immi-gration approach will adjust to this reality.

Take the long view: Changealways wins. The left-right politi-cal divide is essentially aboutwhere people stand with respectto tolerance of social and econom-ic change, giving us progressivesand conservatives. While we aresplit in this country in this regard,those of us who work in technology have tended tohave a deeper understanding of the modern mecha-nisms of change and thus a greater affinity for it. Youcould say we are heavily vested in it.

I think it is helpful to take a longer view of historyand take better stock of the modern era. Compareour world with that of the 1950s, the turn of the cen-tury, or the 18th century. Consider social parameterssuch as voting rights for women or people of color, oreven voting by non-property owners. Consider theshifts in attitude in interracial marriage, in racial seg-regation and slavery. Read through the Wikipediapage on the “Timeline of women’s legal rights” to seehow just a couple of hundred years ago women werenot able to own property.

Your exploration will lead you to one truth: Intime, progressiveness has a good track record ofprevailing. This may come in fits and starts wherewe move one step forward and two steps backward.So, for tech workers who have a high affinity forchange, calm down, take a deep breath and justwait a bit. Things will change. z

Al Hilwa is programdirector of applicationdevelopment software

research at IDC.

Over the years, I have

learned that diversity of

thought in this country is its

greatest source of strength.


Immigration, progress and politics


David Rubinstein is editor-in-chief of SD Times.

50 SD Times January 2017 www.sdtimes.com


Industry WatchBY DAVID RUBINSTEIN

Toys that understand a child’s language andinteract with him or her on a level he or she

can understand. Personal assistants that learn yourroutines and vocalize reminders to complete tasks,or to exercise. Software quality systems that canlearn what is a bug and what is not, or what areasare vulnerable to attack, and then automate aresponse.

Those are just a few examples of where artificialintelligence is today. Where it goes (smart cars,smart security or household robots) is up to theimagination. But 2016 marked the year that artifi-cial intelligence became part of the mainstream,

leaping off the pages of dusty sci-fi books and into the reality ofour everyday lives.

Google, IBM and Microsoftall are building out platforms andproducts for artificial intelli-gence. Google created AlphaGoto defeat a human champion in a

complex game invented in the 4th century BC.IBM, of course, has Watson, which proved that amachine can process and recall data faster thanhumans, but also revealed weakness when it cameto nuance and inference, and sparked an explosionin research into natural language processing.Meanwhile, Microsoft CEO Satya Nadellaannounced at the company’s Ignite conference thatartificial intelligence would be built into the com-pany’s collaboration software for even greaterunderstanding of how people work and whatthey’re working on.

Artificial intelligence is also behind the drive forcompanies to make a digital transformation. Thisphenomenon will result in companies able to com-plete manufacturing with almost no time lost dueto human error or injury. It will allow online retail-ers to create personalized experiences to shoppersby learning their buying habits and offering up in-store specials on items. It will allow organizationsto detect security breaches and contain the dam-age.

And at the heart of this digital transformation isdata—the amount of which is growing exponen-tially each year. With more data to collect andprocess, data administrators and engineers willhave to make their systems smarter to help busi-

ness gain valuable, actionable insights to makebetter decisions.

For digital worker collaboration, two areas havemade this possible: the cloud, and near-always-onconnectivity. Documents, architectural diagrams,pharmaceutical formulae and more can be sharedmore easily than ever. People can start their day byreading e-mail on a WiFi tablet, making Bluetoothcalls on their 4G cell phones on the drive to work(hands-free, of course), and then getting to theoffice and again connecting via WiFi or a land-based server connection.

But this also requires a steadfast commitment tosecurity. As we saw this past year with Yahoo, hack-ers have remained a step ahead of the securityengineers trying to prevent the loss of sensitivedata. That breach exposed some personal informa-tion of Yahoo users. More potentially catastrophicare reports that Russians hacked into DemocraticNational Committee servers to try to impact theUnited States presidential elections. The veryintegrity of nations is at stake here.

Some fear this move toward automation, digitali-zation and robotics, believing the film industry’s por-trayals of robots that start out benign but then gorogue (or their programs are changed by some eviloverlord) and eventually take over society. Some fearthe loss of jobs to machines; we’ve already seen thatin the automotive and other manufacturing indus-tries. Some wonder what will be left for humans todo once machines take over everything. Will we allhave to become computer or data scientists to earn aliving? Buy food? Support our families?

Still others see wonder in these advancements.Imagine a world where your car takes you exactlywhere you want to go, saving you from the stress ofrush-hour traffic or the fear of some wrong-waydriver crashing head-on into your car. Imagine aworld where machines do the housework, cookingand shopping, freeing up time for leisure activity orthe pursuit of other endeavors.

All of this, of course, is still probably decadesaway. But as we witness the very early stages of thedigital transformation of our everyday lives, myhope is that the scientists and engineers creatingthese soon-to-be realities do it from a place ofbeneficence, and don’t let this get out of hand.We’re watching. z

2017: The future starts now

With more data to collect and

process, data administrators

and engineers will have to

make their systems smarter.



http://www.spreadsheetgear.com


http://www.syncfusion.com/SDbigdata