imc summit 2016 breakout - nikita shamgunov - propelling iot innovation with predictive analytics

52
(c) Nikita Shamgunov and MemSQL

Upload: in-memory-computing-summit

Post on 16-Apr-2017

207 views

Category:

Data & Analytics


2 download

TRANSCRIPT

(c) Nikita Shamgunov and MemSQL

Topics• The On-Demand Economy

• From In-Memory Compu8ng to In-Memory Databases

• Renewable Energy and PowerStream

• Demo and Q&A

(c) Nikita Shamgunov and MemSQL

The On-DemandEconomy

(c) Nikita Shamgunov and MemSQL

We live in an on-demand economy

(c) Nikita Shamgunov and MemSQL

Consumers are condi.oned to instant services, like Uber, Stripe, and Airbnb

(c) Nikita Shamgunov and MemSQL

Where does that leave enterprises?

(c) Nikita Shamgunov and MemSQL

Racing to meet internal and external expecta1ons for speed and

personaliza,on

(c) Nikita Shamgunov and MemSQL

Batch processing is the enterprise enemy

(c) Nikita Shamgunov and MemSQL

Enterprises must move from overnight to

Real-&me, intra-day opera&ons

(c) Nikita Shamgunov and MemSQL

Harnessing data in real -me

(c) Nikita Shamgunov and MemSQL

Achieving sub 100 ms latency• Real-'me monitoring and analy'cs on streaming video

• Proac'vely diagnose issues in real-'me

• Deliver be9er viewer experience

(c) Nikita Shamgunov and MemSQL

Massive Ingest AND Analy1cs• Instant accuracy to the latest repin

• Build real-5me analy5c applica5ons

• 1 GB/sec totaling 72 TB/day

(c) Nikita Shamgunov and MemSQL

What is anin-memory database?(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

• Do not (need to) save data on disk

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

• Do not (need to) save data on disk

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

• Do not (need to) save data on disk

• Put the whole dataset in memory

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

• Do not (need to) save data on disk

• Put the whole dataset in memory

(c) Nikita Shamgunov and MemSQL

In-Memory Databases...• Use memory instead of disk

• Do not (need to) save data on disk

• Put the whole dataset in memory

Well, some)mes...

(c) Nikita Shamgunov and MemSQL

Wikipedia says...

In-memory databases primarily rely on main-memory for storage.

(c) Nikita Shamgunov and MemSQL

In-Memory Databases• Are durable to disk (and respect ACID)

(c) Nikita Shamgunov and MemSQL

In-Memory Databases• Are durable to disk (and respect ACID)

• Can spill on disk or pin data in-memory (and take advantage of it)

(c) Nikita Shamgunov and MemSQL

In-Memory Databases• Are durable to disk (and respect ACID)

• Can spill on disk or pin data in-memory (and take advantage of it)

• Tradeoffs are suited to systems with lots of memory

(c) Nikita Shamgunov and MemSQL

In-Memory Databases• Are durable to disk (and respect ACID)

• Can spill on disk or pin data in-memory (and take advantage of it)

• Tradeoffs are suited to systems with lots of memory

• Tend to be distributed systems

(c) Nikita Shamgunov and MemSQL

In-Memory Databases• Are durable to disk (and respect ACID)

• Can spill on disk or pin data in-memory (and take advantage of it)

• Tradeoffs are suited to systems with lots of memory

• Tend to be distributed systems

• Have a different set of boClenecks

(c) Nikita Shamgunov and MemSQL

Bold Claim(c) Nikita Shamgunov and MemSQL

All database workloads will be running on in-memory databases

(c) Nikita Shamgunov and MemSQL

Why?• Memory is ge,ng cheaper (about 40% every year)

(c) Nikita Shamgunov and MemSQL

Why?• Memory is ge,ng cheaper (about 40% every year)

• Cache is the new RAM (RAM is the new disk, disk is the new tape, etc)

(c) Nikita Shamgunov and MemSQL

Why?• Memory is ge,ng cheaper (about 40% every year)

• Cache is the new RAM (RAM is the new disk, disk is the new tape, etc)

• In-memory databases leverage SSD (no random writes)

(c) Nikita Shamgunov and MemSQL

Why?• Memory is ge,ng cheaper (about 40% every year)

• Cache is the new RAM (RAM is the new disk, disk is the new tape, etc)

• In-memory databases leverage SSD (no random writes)

• NVRAM is coming (and could be cheaper than SSD)

(c) Nikita Shamgunov and MemSQL

Why?• Memory is ge,ng cheaper (about 40% every year)

• Cache is the new RAM (RAM is the new disk, disk is the new tape, etc)

• In-memory databases leverage SSD (no random writes)

• NVRAM is coming (and could be cheaper than SSD)

In-memory databases are tuned to modern hardware and modern workloads

(c) Nikita Shamgunov and MemSQL

Renewable Energyin the News

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

IntroducingPowerStream(c) Nikita Shamgunov and MemSQL

MemSQL PowerStreamPredic'ng the global health of wind turbines

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

Streamliner Architecture for PowerStream

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

LiveDemo

(c) Nikita Shamgunov and MemSQL

Demo Sequence• Powerstream user interface

• Showcase largest windfarms

• Real-8me simula8ons

• Witness live opera8ons

• Ease of new pipeline setup

• Ka>a subscrip8on

(c) Nikita Shamgunov and MemSQL

Enabling predic.ve analy.cs• Use exis(ng models from SAS

• Create models in Spark MLlib

• Predic(ve scoring as part of the pipeline

(c) Nikita Shamgunov and MemSQL

From a Real-Time Dashboard to Predic5ve Applica5ons

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL

(c) Nikita Shamgunov and MemSQL