© 2012 IBM Corporation
IBM Cognos OLAP Cubes Re-Visited Juha Teljo
Business Intelligence Solutions Executive
What is IBM’s position on our cube technologies?
Strategic in-memory technologies
§ Dynamic Cubes: High-performance analytics over terabytes of data
§ TM1: Write-back application focused with robust business modeling and rules engine
Invest for on-going customer success
§ PowerCube: Portable, optimized for multi-dimensional analysis
§ Dimensionally Modeled Relational (DMR): Dimensional view over any relational database
IBM’s strategy is the right fit for the right business problem
New in IBM Cognos 10.2 – Dynamic Cubes
Modern and Legacy Sources
Applica4on Sources
3rd Party OLAP Sources
Rela4onal Sources
Dynamic Query Mode
Common Business Model
Classic Query Mode
Scorecards Dashboards
Reports
Ad-‐hoc Query
Analysis & Explora4on
Trend & Sta4s4cal Analysis
What-‐If Analysis
PowerCubes
Open Data Access
OLAP Over Relational
Dimensionally Modeled Relational
Large Enterprise Data Warehouse
Database Aggregates
Dynamic Cubes
§ High performance on high volume star or snowflake schemas in relational sources
§ Powerful in-memory
OLAP cubes § Aggregate aware § Easily-optimized
aggregates § Shareable where
security is shared § Accessible by all IBM
Cognos Interfaces § Included with Cognos
Business Intelligence (no additional cost)
TM1
1. Model & publish
The image cannot be displayed. Your computer may not have enough memory to open the image, or the image may have been corrupted. Restart your computer, and then open the file again. If the red x still appears, you may have to delete the image and then insert it again.
2. Deploy & manage 3. Repor:ng & analy:cs
4. Op:mize
Dynamic Cube Server
Dynamic Cube Logs
CM
Warehouse
Dynamic Cubes Lifecycle
How does PureData Systems for Analytics (Netezza) fit in?
§ Now known as “PureData Systems for Analytics, Powered by Netezza Technology”
§ One of the PureData family of servers – this is the one with Netezza technology and is specifically tuned for BA/AA
§ Netezza is essentially a data warehouse and analytics appliance.
§ It has an MPP (massively parallel processing) database, not columnar as many presume, and specially designed hardware to make analytic DB queries very fast.
§ It can handle high volume, high access speed data – up to 10 petabytes
§ Viewed as simply another Database server to Cognos solutions
§ Not specifically designed for OLAP, but can be leveraged very effectively, especially with Dynamic Cubes, using aggregate tables
§ Should also be excellent for standardized BA reporting on large data volumes and help with DMR analysis
5
System for Analytics
Client Success with Dynamic Cubes Enabling self-service BI at University of Colorado
Enable departments to self-service their own analysis needs
Dramatically improve analytics performance for business users
Optimize resources by saving IT time and effort with easy performance tuning
The need: Faced performance challenges optimizing their relational managed reports centrally delivered to all departments across the 4 campuses of University of Colorado. Report performance was not fast enough, and lacked the ability to enable individual departments to do their own analysis. The solution: In a little over two weeks, delivered a Dynamic Cubes solution that dramatically improved performance of existing reports and also enabled University of Colorado to make a cube available for self-service adhoc analysis for the first time. Real business results: • Department reports that took over a day and required manual manipulation of results, now run in less than 3 seconds without any extra work • Able to make a cube available for ad-hoc analysis, enabling self-service BI for departments • Create dashboards that weren’t previously feasible due to performance challenges – now they run in 5 seconds or less!
“Dynamic Cubes helps us turn Cognos from a packaged reporting engine into a self-service BI engine”
“With Dynamic Cubes, performance will continue to be fast even as our data volumes grow”
— Molly Doyle Assistant Director for IRM
University Information Systems University of Colorado, Office of the President
University of Colorado
What about Size / Performance / Scale / Hardware …?
§ Some of the most common questions but most difficult to answer
§ The right answer changes depending on (to name a few):
– Data volumes – Underlying source data structure – Type of data & complexity (calculations, measures, etc) – Filters & security
§ IBM does not provide benchmark numbers
§ Red books and reference materials do provide general guidelines
§ Deep Dive tech jams are under consideration for Q2/Q3
§ For in-depth discussions look to your TSA or BASA for help
Cognos Dynamic Cubes Redbook
Application Objective Data Structure
Optimal Technology
Notes / Considerations
• Write-back • What-if analysis • High-volatility apps
TM1 Medium data volumes Aggregates on the fly
• High performance analytics • Large data volume • Star / Snowflake schema
Dynamic Cubes
Optimized aggregates Aggregate-aware
• Operational / transactional system
• Consistent performance PowerCubes
Low / medium data volumes Data movement into cube: Latency
Cube groups to manage volume
• Operational / transactional system
• Tightly control latency (cached & non-cached data)
• Tight control over security
DMR (via DQM)
Low / medium data volumes Leverages Framework Manager model
No database aggregate support
Cube Technology Selection – Simple Decision Tree
What about Multiple OLAP technologies?
§ Is it a problem if a customer has more than one cube technology? – NO
§ One Size does not fit all. It is ok to have multiple technologies
§ There is integrated functionality in the platform – Seamless Drill Through – Irrelevant to the end-user what the underlying technology is
§ Professional Report Authoring allow us to mix multiple cube technologies in same report
– Hides data source – Allows end-user to focus on data and results
11
Customers licensing multiple cube technologies
§ PowerCubes, DMR, and Dynamic Cubes come as part of Enterprise BI in Cognos10.2
§ Cognos Insight included for Advanced Business Author upwards
§ TM1 (who need write-back) need to license IBM Cognos Analytics Server
§ IBM Cognos Analytics Server combined with Advanced Business Author upwards allows write-back to TM1 with Cognos Insight
§ For deeper integration - need to license TM1 Contributor in order to include the contributor widgets in Cognos Workspace.
12
How does Dynamic Cubes affect Licensing - PVU
§ Use the sizing papers (Google Dynamic Cube Sizing) and CTP / TSA guidance to determine hardware requirements
§ If PVU entitlements are fully allocated – adding Dynamic Cubes will require more PVU
§ Dynamic Cube capacity counts against PVU capacity just like any other HW Capacity
§ If their machine is powerful enough, they won’t incur additional cost to include Dynamic Cubes
§ If they need a larger configuration, they will need additional PVUs
§ Reference the Pricing and Licensing Center for PVU information
13
Variety
Volume Velocity
Veracity
• Cognos Business Intelligence 10.2 • Dynamic cubes • BigInsights • Netezza • GreenPlum • Paraccel • AsterData • Vectorwise • IBM DB2 Warehouse
• SPSS Predic:ve Analy:cs for data mining and text analy:cs
• InfoSphere Streams • SPSS models
• Real-‐:me scoring • Cognos Real-‐:me Monitoring
• Cognos Consumer Insights • SPSS Data Collec:on Plus the IBM por/olio IBM Content Analy6cs (Filenet) and Customer Experience Suite
• BA Risk PorRolio including recent acquisi:on Varicent
Enhanced for MORE Volume New Analy4cal Models
Purpose-‐built Analy4cs
Area of Investment
Today’s breadth of Big Analytics from IBM Business Analytics
Top Related