Real Time Analytics
Vertica
– A SQL analytic engine
– Built for Speed, Scale and Efficiency
– Supports standard SQL
– Provides rich Analytic functionality and is extensible
– Integrates well with Big Data ecosystem tools
– Runs on premises, in the Cloud, and on Hadoop
What's wrong with this picture?
– SQL ??
– Real-time Analytics ???
– Real-time, continuous load ?
– Real-time, very short response time ?
– Big Data ????
Vertica – Does it scale ???
select GET_COMPLIANCE_STATUS();
Vertica – Does it scale ???(not a fake, believe me…)
select GET_COMPLIANCE_STATUS();
GET_COMPLIANCE_STATUS
--------------------------------------------------------------------------------
Raw Data Size: 2.75PB +/- 0.30PB
License Size : 1.95PB
Utilization : 141%
Audit Time : 2016-09-27 23:59:29.367875+00
Compliance Status : ***** NOTICE OF LICENSE NON-COMPLIANCE *****
Continued use of this database is in violation of the current license agreement.
Maximum licensed raw data size: 1.95PB
Current raw data size: 2.75PB
License utilization: 141%
IMMEDIATE ACTION IS REQUIRED, PLEASE CONTACT VERTICA
Vertica – Is it really fast ?
– Trillion Row Qlik-on-Vertica Dashboard
– https://www.youtube.com/watch?v=ZnMDeg8V2sg
Vertica – Is it so simple ?
– HPE Vertica and Qlik Direct Discovery: A Technical Exploration
– https://community.dev.hpe.com/t5/Vertica-Knowledge-Base/HPE-Vertica-and-Qlik-Direct-Discovery-A-Technical-Exploration/ta-p/234332
Vertica – Is it so simple ?
– No !
– HPE Vertica and Qlik Direct Discovery: A Technical Exploration
– Implementation Methods
– Fact and dimension tables in-memory. Most applications are created using this approach. However, this paper does not cover the all-in-memory option because it is not suitable for big data (such as a few billion rows of fact data) and requires too much memory.
– Fact and dimension tables in Direct Discovery (regular star schema).
– BFFT (big flat fact table) in Direct Discovery. There are no dimension tables with BFFT.
– Fact tables in Direct Discovery and dimensions in memory.
– Multiple fact tables in Direct Discovery. This is not generally recommended because of complex design considerations.
Vertica @ Nimble Storage
10
Changing the game with the Internet of (Powerful) Things
InfoSight
Nimble Storage – Some metrics
– >7,500 customers
– millions of virtual objects under continuous monitoring
– collected per day
– Database Characteristics
– Raw Data : 550TB - Disk: 200 TB - On Nimble: 100 TB
– 350K selects per day
– 60K inserts/deletes per day
– Configuration
– 2 Vertica clusters – 2x8 servers – 2x8x54 cores – Nimble Storage instead of DAS
>250 billion sensor values
>2 billion log events
>100 million configuration variables
More on Vertica by Nimble Storage
– https://my.vertica.com/wp-content/uploads/2016/09/B10823_10823_Presentation_2.pdf
– From Vertica Big Data Conference 2016 : https://my.vertica.com/big-data-conference-2016/
Vertica @ Criteo
14
Hadoop for Primary Storage
and MapReduce
Cascading, Scalding and
Hive for Data Transformation
Hive and Vertica for
Data Warehousing
Tableau and ROLAP Cube
for Structured Data Access
Vizatra for speed
The analytics stack at Criteo
More on Vizatra+Vertica by Criteo
–SBTB FinagleCon 2015: Justin Coffey, Presenting Vizatra – YouTube
–https://www.youtube.com/watch?v=uXmEhSFzNLs
More on Vertica
Vertica analytics platform
Fast
Boost performance by 500% or more
Scalable
Handles huge workloads at high speeds
Standard
No need to learn new languages or add complexity
Costs
Significantly lower cost over legacy platforms
18
About Vertica
Massively Parallel Processing
– Shared Nothing
– Elastic scale-out architecture
– Built-in high availability
– Commodity Hardware
– Easy setup and administration
– And more …
Client Network
Private Data Network
20 TB 20 TB 20 TB
Node 1 2 x 12 Cores 128+GB RAM
Node 2 2 x 12 Cores 128+GB RAM
Node 3 2 x 12 Cores 128+GB RAM
Core Vertica TechnologyBuilt for performance and scale
20
my.vertica.com
–Download Vertica Community Edition on my.vertica.com
–Up to 1 TB and 3 nodes
21