5 crucial considerations for big data adoption

18
5 Crucial Considerations for Big Data Adoption: Qubole On AWS vs. In-House Infrastructure Deployment

Upload: qubole

Post on 06-Apr-2017

3.539 views

Category:

Data & Analytics


1 download

TRANSCRIPT

Page 1: 5 Crucial Considerations for Big data adoption

5 Crucial Considerations for Big Data Adoption:

Qubole On AWS vs. In-House Infrastructure Deployment

Page 2: 5 Crucial Considerations for Big data adoption

PRESENTED BY:

Page 3: 5 Crucial Considerations for Big data adoption

Only 13% of organizations achieve full-scale production for their in-house big data implementations.

13%

RISK

Page 4: 5 Crucial Considerations for Big data adoption

ONLY 27% OF EXECUTIVES DESCRIBED THEIR IN-HOUSE BIG DATA INITIATIVES AS SUCCESSFUL.

Page 5: 5 Crucial Considerations for Big data adoption

Boosting time-to-value with a big data project is crucial to keeping up in a fast-paced market. Consider the following factors to streamline big data adoption.

Page 6: 5 Crucial Considerations for Big data adoption

Time to deployment

Average reported in-house infrastructure project build

times (not production)6-9 months.

6-9

TIME VALUE

Page 7: 5 Crucial Considerations for Big data adoption

*QUBOLE AVERAGE USER TIME TO FIRST PRODUCTION QUERY = 2.8 DAYS

Page 8: 5 Crucial Considerations for Big data adoption

Datasets will grow rapidly which means infrastructure will need to grow too.

LONG TERM SCALABILITY

Page 9: 5 Crucial Considerations for Big data adoption

ON-PREMISE EXPANSION CAN TAKE WEEKS OR MONTHS, SO PLAN TO SCALE SEVERAL MONTHS OUT WHICH MEANS PROCURING ADDITIONAL HARDWARE.

With Qubole on Amazon Web Services, the average time it takes to spin up a 200 node cluster is 4 minutes.

200 NODES: 4 MINUTES

Page 10: 5 Crucial Considerations for Big data adoption

THERE ARE 100+ PROJECTS WITHIN THE HADOOP ECOSYSTEM

Each big data tool has a specific use case and requires specialized skills to use.

Big data vendors offer varying levels of support to reduce the skills gap.

ASSEMBLY REQUIRED:WILL HADOOP CONSUME YOUR COMPANY?

Page 11: 5 Crucial Considerations for Big data adoption

On-premise distributions require 5-10 staff members to manage large clusters

(1000+ nodes).

Qubole customer: a single IT manager can manage all projects regardless of

size or cluster count.

Page 12: 5 Crucial Considerations for Big data adoption

INFRASTRUCTUREMANAGEMENT REQUIREMENTS

ON-PREMISE:cluster sizing, configuration management, health and performance monitoring, resource utilization and control, project management

Page 13: 5 Crucial Considerations for Big data adoption

QUBOLE ON AWS:project management, vendor coordination

Page 14: 5 Crucial Considerations for Big data adoption

THE MORE PEOPLE THAT HAVE ACCESS TO DATA, THE MORE USEFUL IT IS.

ACCESSIBILITY

Page 15: 5 Crucial Considerations for Big data adoption

Ease of accessibility varies by vendor. Managed services offer greater

accessibility to non-IT teams.

Common Struggles: Complex Tools, Strain on IT Resources,Teams need different tools,Training takes significant time.

Page 16: 5 Crucial Considerations for Big data adoption

63%

57% of organizations cite skills gap as a major inhibitor to Hadoop adoption.

*63% of Qubole users report little or no training was required for analysts

to start analysing data.

57%

Page 17: 5 Crucial Considerations for Big data adoption

Interested in learning how the cloud can help

you derive faster time to value from big data?

Watch this webinar from Forrester Research.

Watch the Webinar

Page 18: 5 Crucial Considerations for Big data adoption

*SOURCE: Qubole Customer Survey April 2015

https://www.capgemini-consulting.com/resource-file-access/resource/pdf/c

racking_the_data_conundrum-big_data_pov_13-1-15_v2.pdf

http://www.gartner.com/newsroom/id/3051717

http://dataconomy.com/the-building-blocks-of-a-data-driven-enterprise/?utm_content=buffer9d010&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

https://hadoopecosystemtable.github.io/file:///Users/a97thFloor/Downloads/MapR%20TCO%20Model%20-%20Hadoop%2020%20node%20TCO%20Template%20[2015-07-02%20422pm].pdf