event driven real time analytics

46
T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com Jon Mead, Rittman Mead September 2012 Event Driven Real Time Analytics Sunday, 30 September 12

Upload: alina-anchidin

Post on 14-Apr-2018

227 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 1/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Jon Mead, Rittman MeadSeptember 2012

Event Driven Real Time Analytics

Sunday, 30 September 12

Page 2: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 2/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Introductions

• Jon Mead

‣CEO/co-founder of...

•Rittman Mead Consulting

‣Oracle BI & DW Consultancy

‣Gold Partner 

Long(est) running Oracle BI blog‣ Annual BI Forum

‣OBIEE Oracle Press book

•Customer-facing FTSE listed

‣UK based and leading

‣ Internet based‣Retail based

Sunday, 30 September 12

Page 3: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 3/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

The point of thispresentation isto give you an

idea of how toapproach a realtime eventdriven BIsystem using

Oracle's currenttoolset.

Agenda

•Understanding the Project

‣ Legacy architecture

‣Proposed architecture

‣Reporting requirements

•Technical Infrastructure

‣Hardware and Software

Data Warehouse Architecture‣ Adopting the Oracle reference architecture for real

time

•Design Challenges

‣De-queuing

•Operational

‣ODI Logging‣Multi-threading and scalability

•Further thoughts

‣Middleware or memory based applications

Sunday, 30 September 12

Page 4: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 4/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Understanding the Project

Sunday, 30 September 12

Page 5: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 5/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Business Goal

• Part of a major re-architecture program

• Covering ERP, CRM and BI

Driver: single view of customer 

Delivered by: channel consolidationinto single enterprise data warehouse

• Data migration• Enterprise Architecture

• Enterprise Service Bus

• Real-time reporting• Legacy reporting

• BAU reporting

• Revenue and Profit• Liability and risk

• Up/cross-sell

Sunday, 30 September 12

Page 6: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 6/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Legacy Architecture

• Legacyarchitectureconsisted of twocompletelyseparate

systems

• Retail storedshop basedtransactions

• Online storedtransactionsgenerated online

Retail trading systems

Online trading systems

Online trading systems

Online trading systems

24 hour batch (DTS)

Retail Data Warehouse

SQL Server 2005

6TB

24 hour batch (DTS)

Online Data Warehouse

SQL Server 2008

3TB

Retail trading systems

Retail trading systems

Sunday, 30 September 12

Page 7: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 7/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Proposed Architecture

Online trading systems

Online trading systems

Enterprise Architecture

Online trading systems

Real-time feed

TIBCO QueueReal Time Data

Warehouse

Exadata

Real-time feed

Real-time feed -

transactions

Real-time feed -

reference data

DR

Exadata

ODI

ODI

Retail Data Warehouse

SQL Server 2005

6TB

Online Data Warehouse

SQL Server 2008

3TB

   D  a   t  a    m   i  g    r  a   t   i  o

   n 

  -   o   n

  c  e   o   f   f   O

   D   I

   D   a   t   a   m   i   g   r   a   t   i   o   n

  -   o   n   c   e   o   f   f

      O      D      I

Retail trading systems

Retail trading systems

Retail trading systems

Sunday, 30 September 12

Page 8: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 8/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Proposed Architecture

Online trading systems

Online trading systems

Enterprise Architecture

Online trading systems

Real-time feed

TIBCO QueueReal Time Data

Warehouse

Exadata

Real-time feed

Real-time feed -

transactions

Real-time feed -

reference data

DR

Exadata

ODI

ODI

Retail Data Warehouse

SQL Server 2005

6TB

Online Data Warehouse

SQL Server 2008

3TB

   D  a   t  a    m   i  g    r  a   t   i  o

   n 

  -   o   n

  c  e   o   f   f   O

   D   I

   D   a   t   a   m   i   g   r   a   t   i   o   n

  -   o   n   c   e   o   f   f

      O      D      I

Retail trading systems

Retail trading systems

Retail trading systems

Current state to future

state includes a data

migration

Sunday, 30 September 12

Page 9: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 9/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Need tounderstand thedifferent driversfor each of these needs and

the valueprovided by realtime reportingduring the

running of hightransactionevents

Reporting Requirements

•Real time monitoring

‣Risk and liability‣Profit and loss

• Analytic reporting

‣Consolidated analytics

‣ Legacy reporting

•Operational reporting

‣Detail level

‣Support analytical reports

‣Drill through

Sunday, 30 September 12

Page 10: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 10/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Technical Infrastructure

Sunday, 30 September 12

Page 11: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 11/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Volumetrics

• Initially processing data from 2500 shops, scaling to capacity

‣ 8 TB of migrated data‣Processing 1.8 million transactions a day

‣Processing 4,000 reference data items a day

‣ Approximately 9 million transaction rows being processed a day

‣ All transactions read from a TIBCO queue

‣ Approximately 200,000 reference data changes a day

‣ 30,834 transaction processing cycles a day (one every ~2.8s)

‣ 2,701 reference data cycles a day

‣ 680,000 recalculations a day

•Online transactions will follow

‣ 2 million transactions a day‣Comparable downstream figures

Sunday, 30 September 12

Page 12: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 12/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Exadata and ODI

•Standard X2-2 Quarter Rack

‣ 2 compute nodes‣ All databases split across the nodes

•The storage is configure in dual redundancy mode to offer up about 9TB of usablespace, however, we use a couple of TB of that for backups and archive redo logs

•The flash storage has been set up as 250GB on each node as a local cache and 110GBfrom each being used to provide a 160GB flash disc.

•The database version is 11.2.0.3 and the client have the tuning and diagnostics packand Heterogeneous Services on top of the usual Exadata software set.

‣Both the 11.2.0.2 and 11.2.0.3 Oracle Homes still exists.

•ODI Agents for UAT and PROD running off Node 2

•Running up to 30 ODI Agents for PROD to get the speed to read off the TIBCO queues.Each agent running with 512MB with the calling agent running 1GB.

Sunday, 30 September 12

Page 13: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 13/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Exadata and ODI

NODE 1

DEV01

DEV02

NFT

PROD

NODE 2

Compute Nodes

OracleDatabases

ODI AGENTS

DEV/ NFT

ODI AGENTS

PROD

UAT

ODIInstalls

PROD ODI

WORK

SCHEMA

UAT

REPOSITORY

PROD

REPOSITORY

UAT ODI

WORK

SCHEMA

UAT

SSD

Sunday, 30 September 12

Page 14: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 14/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Exadata and ODI

NODE 1

DEV01

DEV02

NFT

PROD

NODE 2

Compute Nodes

OracleDatabases

ODI AGENTS

DEV/ NFT

ODI AGENTS

PROD

UAT

ODIInstalls

PROD ODI

WORK

SCHEMA

UAT

REPOSITORY

PROD

REPOSITORY

UAT ODI

WORK

SCHEMA

UAT

SSD

Currently only one Exadata

server available, so shared

platform

Sunday, 30 September 12

Page 15: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 15/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Data Warehouse Architectur e

Sunday, 30 September 12

Page 16: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 16/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Key Drivers

•Part of integrated Enterprise Architecture

•The enterprise data model was designedand developed in Enterprise Architect by themiddleware architects

•The architects wanted to base the approach

on the Oracle Reference Data Warehousearchitecture

•There were different reporting needs for real time and business as usual reporting

•Write performance was likely to be as big afactor as read performanc

Sunday, 30 September 12

Page 17: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 17/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Oracle Reference Architecture

•Simplified view of Oracle’s

Data Warehouse Reference Architecture

•Enterprise Architecture wasXML based

Active Data Warehouse

Staging

Foundation

 O

B I   E E 

PerfromanceAudit and

Reconciliation

Analysis

Operationaland real-

time

Enterprise Architecture

Sunday, 30 September 12

Page 18: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 18/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Oracle Reference Architecture

•Simplified view of Oracle’s

Data Warehouse Reference Architecture

•Enterprise Architecture wasXML based

Active Data Warehouse

Staging

Foundation

 O

B I   E E 

PerfromanceAudit and

Reconciliation

Analysis

Operationaland real-

time

Enterprise Architecture

One of design drivers was that the

foundation layer reflected the

enterprise data model

Sunday, 30 September 12

Page 19: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 19/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Limitations

•The ODS would reflect the

enterprise architecture‣Non-database centric view

•Considerable processing to getdata into ODS

‣Data processing from Stagingto Foundation was too

complex to support SLAs•Performance layer also to

reflect existing more datawarehouse structures

Sunday, 30 September 12

Page 20: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 20/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Limitations

•The ODS would reflect the

enterprise architecture‣Non-database centric view

•Considerable processing to getdata into ODS

‣Data processing from Stagingto Foundation was too

complex to support SLAs•Performance layer also to

reflect existing more datawarehouse structures

Result was non-performant and

unusable structures for real-time

reporting

Sunday, 30 September 12

Page 21: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 21/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Alternative Architecture

• Split the processing between real-time and BAU

‣Process 1: Staging to Performance (real-time)

‣Process 2: Staging to ODS to Foundation

ODI

STG_xxx

STG_xxx

STG_xxx

STG_CTL

Staging

     T     I     B      C      O

ODI

Microbatch ETL

BETBETSLIPDecompositionand Aggregate

tables

PerformanceFoundation

ODI Real time ETL

Real timeETL

      O     B     I     E     E

SQL

SQL

Real time

query

Near realtime

BETBETSLIPDimension and

fact tables

ODI

Microbatch ETL

SQL

Near realtime

BETBETSLIP

3NF tablesBETBETSLIP

3NF tables

• Independentcontrol of either process

• Mechanism tohandle peaks indata

• Needed to ensureconsistencybetweenprocesses

Sunday, 30 September 12

Page 22: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 22/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Design Challenges

Sunday, 30 September 12

Page 23: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 23/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

De-queuing

•Concern that ODI would not be able to de-queue

‣ A lot of fluctuations, depending on events

•XML messages were verbose

‣ Large amount of processing time for each batch of messages

•Scalability provided by creating more agents

‣What would the limitations be in terms of RAM

‣What would the limitations be in terms of connections

‣What would the limitations be in terms of management

Sunday, 30 September 12

Page 24: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 24/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

2 Queues

•Queue are unstructured so

the data can arrive in anyorder 

‣Difficulty of processingbusiness logic

‣ Timestamps not alwaysaccurate

‣Keys not always present

•One of most challengingareas of the project

‣Often need to do manuallookup of keys

Transactions

STG_CTL

     T     I     B      C      O

Real timeETL

Referencedata

Recycle

STG_xxx

STG_xxx

STG_xxx

Solution: recycle

mechanism

Sunday, 30 September 12

Page 25: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 25/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

High number of writes

•The real-time write process generated a very high number of writes

•Exadata optimised for bulk reads•Contention for REDO logs (see also the ODI Logging)

•Exadata configured for more of an OLTP system than Data Warehousing system

‣However both share the same server 

•Resolution: lots of work by the DBAs to optimise database configuration

Isn't this a little bit like an

OLTP system?

Sunday, 30 September 12

Page 26: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 26/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Operational Challenges

Sunday, 30 September 12

Page 27: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 27/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI Parser 

•XML Parser 

‣ The XML data definition files were dynamically generated.‣ The current version of the XML Parser does not do a double pass of the definition file

‣ Any referenced complex definitions needed to be defined in the order they wereaccessed

‣ The software generating the XML definition files did not do this

•Resolution

‣Build a Java program to re-parse the XML data definition file and output a correctlyordered version

‣ This is a once per release process

‣ This behaviour if fixed in 11.1.1.7 of ODI (I think)

Sunday, 30 September 12

Page 28: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 28/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

De-queuing Performance

•ODI struggled to keep up with, or fell behind the queue at peak times

‣Volumes of messages were not regular 

•We also found agents failing

‣Hence we needed a resumption mechanism

Sunday, 30 September 12

Page 29: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 29/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Scaling agents

•Because of the failing agents, we couldn’t just

increase their number 

•Set up parent agents

‣One for each queue

‣One for monitoring and maintenance scripts

•Each parent agent ran a number of child agents

‣Each child agent was actually two agents

‣Second agent acted as redundancy

‣ Agents killed after 50 executions

P1 P2 M&M

A1 A2 A3 A4 C6

A3 A4

C3 C4

A1 A2

A1 A2

C1 C2

Q1 Q2

Sunday, 30 September 12

Page 30: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 30/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Connections:

Every time anODI agent readfrom thequeue a newconnectionwas createdand destroyed.

There didn’tseem to beany pooling.

Impacts of Multiple Agents

•Memory

‣Parent agent 1024MB‣Child agent 512M

•Total number of agents used

‣ 3 parents

‣ 18 child (primary) and 18 child (secondary)

‣ Total: 39 = 21504MB (approx 21GB)

•However we didn’t get anywhere near linear scaling

‣Max TPS = 176

‣Max queue TPS = 480

•Second option is to increase the number of queues

‣Split by functional area

Sunday, 30 September 12

Page 31: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 31/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI Logging

•ODI Logging

‣ The ODI processes create 900GBof log files a day

‣ODI logging needs high IOPS

‣Exadata, by default not allocatingenough IOPS resource to the ODIlogging

‣ODI logging then becomes alimiting factor on the databaseperformance

‣ Target is SNP_SESS_TASK_LOG

‣ Log writer process cannot keepup

‣Number of active processes meanthe database will be performingas hard as it can and more activitywill slow everything down.

!

Sunday, 30 September 12

Page 32: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 32/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI Logging

•ODI Logging

‣ The ODI processes create 900GBof log files a day

‣ODI logging needs high IOPS

‣Exadata, by default not allocatingenough IOPS resource to the ODIlogging

‣ODI logging then becomes alimiting factor on the databaseperformance

‣ Target is SNP_SESS_TASK_LOG

‣ Log writer process cannot keepup

‣Number of active processes meanthe database will be performingas hard as it can and more activitywill slow everything down.

!

ODI does the same logging but the volumepreserved is reduced with lower levels of logging.

So in fact, lower levels of logging could be more

IO demanding as more data is deleted.

Sunday, 30 September 12

Page 33: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 33/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

SNP_SESS_TASK_LOG

•The most demanding

SQL statement on thesystem is and always hasbeen the update toSNP_SESS_TASK_LOG

•This table holds 3 CLOB

columns. The update is"lazy", all columns areupdated each time. Thuseach update canpotentially update:

‣ the table

three clob indexes‣ three clob tables

Sunday, 30 September 12

Page 34: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 34/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI Logging Impact

• Initially the system was

throttled on theSNP_SESS_TASK_LOG.

•ODI IOPS demand wasmaxing out the physical discIOPS capability of the box

•MoveSNP_SESS_TASK_LOG andSNP_SESS_TASK to a new ASM diskgroup created fromthe SSD storage in Exadata

!

Sunday, 30 September 12

Page 35: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 35/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI Logging Impact

• Initially the system was

throttled on theSNP_SESS_TASK_LOG.

•ODI IOPS demand wasmaxing out the physical discIOPS capability of the box

•MoveSNP_SESS_TASK_LOG andSNP_SESS_TASK to a new ASM diskgroup created fromthe SSD storage in Exadata

!

Simple solution for the ODI

Logging problem is to movethe database to another

server

Sunday, 30 September 12

Page 36: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 36/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI temporary tables

•The ODI real-time processing

creating a large number of I$ andother internal tables

•Once the processing around theseis complete, they are put theRecycle Bin

•The Recycle Bin become either 

full or unmanageable

Sunday, 30 September 12

Page 37: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 37/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

ODI temporary tables

•The ODI real-time processing

creating a large number of I$ andother internal tables

•Once the processing around theseis complete, they are put theRecycle Bin

•The Recycle Bin become either 

full or unmanageable

There is a wider issue here

that affects scalability of the

whole solution, discussed innext slides

Sunday, 30 September 12

Page 38: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 38/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Future Challenges

Sunday, 30 September 12

Page 39: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 39/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Accessing datain memoryreduces the I/Oreading activity

when queryingthe data whichprovides faster and more

predictableperformancethan disk

Scalability

•The main bottleneck we are experiencing is I/O

‣High number of writes to the database‣ODI $ internal tables

‣ODI logging

•We should address this problem by making better use of memory

•We also have a constraint on the amount of memory Exadata can provide the agents

‣ Any allocated memory has the opportunity costof not be used by the database

•We should also explore other ‘logical’ approachesto solving this problem

Sunday, 30 September 12

Page 40: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 40/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Process Flexibility

•The source system splits the data into real

time and batch‣ESB provides 2 separate queues

•The XSD is the same for both queues

•The processing for both queues isconstantly running in a loop

‣ The batch queue is much larger than the

real time queue.‣ The foundation layer requires data from

both

•The data from each queue lands in thesame stage tables partitioned by thequeue name

•Entire process controlled by maintainingBATCH_IDs

SRC SystemsFeed

Non Critical Data Real Time Data

 

Stage Schema

DQProcess

DQProcess

Event StageTables

STG_ Tables

 

Performance Schema

 

Foundation Schema

Reporting Tables

ODS LoadProcessing

Real Timeprocessing

Foundation LayerTables

CTL_EVENTbatch_idbatch_typeODS_processedRTF_processedSTG_processed

CDC ProcessCDC Process

Sunday, 30 September 12

Page 41: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 41/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

In-Memory Processing

•Will require re-writing of the Knowledge Modules

‣Should also persist connections

•Option 1: Remove the writes to the $ tables and attempt to do moreoperations on-the-fly

‣Potential loss of audit trail and reconciliation points

‣Currently all outer joins are materialised

‣Will need to perform

•Option 2: Use In-Memory database?

Sunday, 30 September 12

Page 42: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 42/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

In-Memory Processing

•Will require re-writing of the Knowledge Modules

‣Should also persist connections

•Option 1: Remove the writes to the $ tables and attempt to do moreoperations on-the-fly

‣Potential loss of audit trail and reconciliation points

‣Currently all outer joins are materialised

‣Will need to perform

•Option 2: Use In-Memory database?

Exalytics?

Sunday, 30 September 12

Page 43: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 43/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Conclusion

Sunday, 30 September 12

Page 44: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 44/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

The objectivesof the projectwhereachieved. Theresulting datais being usedon a dailybasis andproving

significantvalue to theorganisation

Conclusion

•The Oracle Reference Data Warehouse architect can

support real time event driven ETL, however it mayneed modifications

• IDO has some rough edges and kinks that need to beironed out for it to act at this kind of enterprise level

•Don’t underestimate the effort of doing a data migration

• Its important to understand the implications anddifferences of middleware centric data models andprocessing compared with databases centric ones

Sunday, 30 September 12

Page 45: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 45/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Questions?

Sunday, 30 September 12

Page 46: Event Driven Real Time Analytics

7/30/2019 Event Driven Real Time Analytics

http://slidepdf.com/reader/full/event-driven-real-time-analytics 46/46

T : +44 (0) 8446 697 995 or (888) 631 1410 (USA) E : [email protected] W: www.rittmanmead.com

Jon Mead, Rittman MeadSeptember 2012

Event Driven Real Time Analytics