IBM COGNOS TM 1 V 10 New Innovations

The 2011 BI Survey 10 ranked IBM Cognos TM 1 … • Goal achievement • Self-service The Planning Peer Group BARC BI Survey 10 • Goal achievement • Self-service • Suitability The Database Peer Group BARC BI Survey 10 Cognos TM 1 10 PART OF COGNOS ENTERPRISE • Improved, flexible solution design environment • Easily spans analytics and planning requirements • Unrivaled performance at high user and data scale 2 © 2012 IBM Corporation

Cognos Enterprise Performance Management: A Revolution with TM 1 10 w: ng e N ypi tot o r P Performance Reporting and Scorecarding Analytics Customer Service Sales Expense Planning & Control Operations Workforce Planning Human Resources Product Development Marketing Capital Planning Finance Strategic Finance Enterprise Planning and Analysis IT Profitability Modeling New So Des lution Dep ign & loym ent High Performance Platform 3 © 2012 IBM Corporation

Flexible Solution Design Environment Simplifies design for performance management solutions Empowers MORE planning application development, more prototyping Easily wires Cognos Insight “personal analysis” into enterprise solutions Spans Address Broad Range of Performance Management Challenges 4 © 2012 IBM Corporation

Fast, Easy Modeling Guided modeling steps – define dimensions – build cubes – define rules – easy linking paradigm Widespread use of drag and drop and visual composition Easy transfer of applications from development to test to production environments Empowering more users to model, reduced time to value 5 © 2012 IBM Corporation

Simple Application Deployment Support for multiple application types, clients and connections – Workflow/Non-Workflow • Approval Hierarchy/Continuous – Client support • Web, Contributor, Cognos Insight – Connection modes • Centralized or Distributed Application Deployment – Development or Production – Cognos Connection or 3 rd Party Portal Distributed Cognos TM 1 Server Centralized Cognos TM 1 Server Unrivaled flexibility and performance at high user and data scale 6 © 2012 IBM Corporation

Takes insight into action with Cognos Insight Users work with their data of choice, on their desktop, without IT Discover trends and explore scenarios via speed-ofthought interactions Share findings & results within a workgroup or throughout the enterprise New Users Connected to PM Processes 7 © 2012 IBM Corporation

New, visually rich and interactive “canvas” for smarter contribution Analyze – Prototype situations, plans – Work with existing or pull in new data – Perform what-if scenarios, associative analysis, profitability analysis – Customize views and “explore points” Prototype a plan – Create new version, shortcuts, views, charts – Create or add to dashboards – Add Instructions Share and / or Initiate new workflow – Promote insights in a tailored dashboard – Collect new information via workflow Foundation of an agile planning and analytic platform 8 © 2012 IBM Corporation

User Experience Flexibility Desktop Excel-Based Customer Service Sales Marketing Operations Product Development Managed Collaboration Finance Human Resources IT Span All Users and User Experience Needs 9 © 2012 IBM Corporation

Proven Analytics Engine Distributed Cognos TM 1 Server – Improved WAN performance – Better user scale Built on the high-performance, stable 9. 5. 2 Server Engine First-in-line with parallel interaction More tools to accelerate ETL performance Multi-server management console Distributed Cognos TM 1 Server +20% Contributor performance improvement Delivers Quality, Scale, and Performance 10 © 2012 IBM Corporation

Providing a full range Business Analytics Integration with IBM's Business Analytics Portfolio Advances insight with: Business Intelligence Predictive and Advanced Analytics Risk Analytics 11 © 2012 IBM Corporation

TM 1 Ecosystem – Cognos BI with Dynamic Query Mode � The first customer test application had an in-memory footprint of 12 GB in Cognos TM 1. Testing included both reporting and interactive exploration gestures, focusing on one dimension within the Cognos TM 1 cube, comprising 150, 000 members. This customer application was developed to manage reporting and analysis for a complex manufacturing environment and included 12 different reports which were run interactively. Figure 1, below, shows the total average report execution times across 12 reports in the complex manufacturing application test suite. Only the first report execution time is measured, as this more closely emulates the “real life” experience of a user running a report once in a given user session. This avoids use of the powerful caching facilities of the Cognos Business Intelligence Platform. The performance of subsequent individual or concurrent report runs would benefit from the use of caching and would therefore show even better performance for subsequent report runs than those shown in this test. 12 © 2012 IBM Corporation

TM 1 Ecosystem – Cognos BI with Dynamic Query Mode � In addition to normal reporting use cases, the complex manufacturing application test suite included a series of interactive data exploration gestures. Figure 2, below, shows the total average response times for individual user gestures that would be typical of a user exploring a Cognos TM 1 application. Among the gestures shown in Figure 2, gestures 4 through 8 all require a significant amount of processing, and it is these process-heavy gestures that enjoy the greatest performance improvements. The improvement in the performance of gestures 4 through 8 in Figure 2 is directly attributable to the superior query and metadata work planning and advanced interactions with Cognos TM 1 9. 5. 2, inherent to the Cognos Dynamic Query mode. 13 © 2012 IBM Corporation

TM 1 Ecosystem – Cognos BI with Dynamic Query Mode � The second customer test application had an in-memory footprint of 18 GB in Cognos TM 1. Testing included both reporting and interactive exploration gestures. However, this test shows the performance comparison between compatible MDX and Cognos Dynamic Query for an application under a concurrent interactive user load equivalent to the load expected for a population of 30 concurrent users (1 concurrent user = 10 active users = 100 named users). This test case is based upon a customer’s production Sales Reporting application and uses a Cognos TM 1 cube that contains 5 dimensions and a possible 9, 377, 523, 949, 248 (9. 3 trillion) total intersections. Figure 3, below, shows the dimensional and cube characteristics of the Sales Reporting application used for testing Cognos Dynamic Query with Cognos TM 1 9. 5. 2. 14 © 2012 IBM Corporation

TM 1 Ecosystem – Cognos BI with Dynamic Query Mode � The test scenarios used for this round of concurrent user testing are typical of any product-driven Sales Reporting application, including the interactive running of pre-defined reports as well as typical userdriven interactive analysis gestures. A large cross-section of reports showing sales by Product, Geography, Time, and Plan, at varying levels of each hierarchy, were used to show the impact of varying data density and aggregation processing on Cognos Dynamic Query performance. � Figure 4 shows the total average execution times across all reports in the Sales application test suite. Only the first report execution time is measured for each concurrent request, as this more closely emulates the “real life” experience of a user running a report once in a given user session. This avoids use of the powerful caching facilities inherent to the Cognos Business Intelligence Platform. The performance of subsequent individual or concurrent report runs would benefit from the use of caching. 15 © 2012 IBM Corporation

TM 1 Ecosystem – Data Loads from Netezza to TM 1 � Load tests performed in the Boeblingen Lab. Data loads (via ODBC) can be parallelised with almost linear scale per thread. Some overhead was noticed within TM 1, but none on Netezza. 16 # of Threads # of Rows # of Measures Loadtime Rows/ second 1 115 Million 2 45 mins 85‘ 185 1 115 Million 5 60 mins 159‘ 722 8 1 Billion 5 77 mins 1‘ 082‘ 251 © 2012 IBM Corporation