Torben Noer WW Business Analytics Tech Sales manager Addedo TM1 und BI User Group Zürich, 28 th Nov 2017

Transcription

1 Got Cognos Analytics and IBM Planning Analytics? Tips for Improving Your ROI Torben Noer WW Business Analytics Tech Sales manager Addedo TM1 und BI User Group Zürich, 28 th Nov 2017

2 Please Note IBM s statements regarding its plans, directions, and intent are subject to change or withdrawal without notice at IBM s sole discretion. Information regarding potential future products is intended to outline our general product direction and it should not be relied on in making a purchasing decision. The information mentioned regarding potential future products is not a commitment, promise, or legal obligation to deliver any material, code or functionality. Information about potential future products may not be incorporated into any contract. The development, release, and timing of any future features or functionality described for our products remains at our sole discretion. Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon many factors, including considerations such as the amount of multiprogramming in the user s job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve results similar to those stated here. IBM Analytics Nov 2017

3 The Basics Upgrade to reimagined user experiences Cognos Business Intelligence v10 to Cognos Analytics (CA) v11 Cognos TM1 v10 to Planning Analytics (PA) v2.0 PA is a data source to CA Managed & self-service CA reports, dashboards and stories on your PA data CA authoring is the same as with other dimensional (MDX) sources Various authentication integration options available 3

4 Upgrading 4

5 Upgrading to CA / PA from BI v10 / TM1 v10 No report left behind All v10 reports upgrade to CA No need to change reports Reminder: CQM to TM1 is not supported in CA Legacy capabilities still available if needed e.g., Framework Manager, Cognos Workspace, Analysis Studio, etc. Legacy integration of TM1 websheets and cube views in Cognos Workspace is not currently working in CA v11 fix planned for a future version of PA (no fix planned for TM x) 5

6 What to do in Cognos Analytics when Upgrading from TM1 to Planning Analytics Upgrade CA to version or higher In Cognos Administration within each existing data source Add a Planning Analytics connection Removing existing TM1 connection Republish the packages Confirm reports run successfully 6

7 Benefits of CA s PA connection over TM1 connection Support for PA Hierarchies Arguably the most important new feature in PA 2.0 Smaller more flexible PA models Faster queries, less memory Utilizes PA s REST API Easier setup More flexible integration Supports localized captions No longer need to uses aliases to present a cube in different languages 7

8 IBM Performance Lab Benchmarking CA s PA connection vs TM1 connection Single user workload Most reports had comparable performance More reports were faster than slower Many optimizations now on by default Other performance improvements thanks to REST API 16+ simultaneous users workload Known degradation in version Spike in TM1 server memory Fixed in version

9 Aliases Present users with more familiar names for members (elements) e.g., the alias for product type code is Helmets Legacy use case for localizing cubes New localized caption approach is better Very different handling in CA for PA compared to TM1 Aliases are now available in CA as attributes / properties Aliases cannot be assigned as captions of the design language when publishing a package Workaround is to use Turbo Integrator (TI) to set captions by alias values set the Caption member in the }ElementAttributes_<dim> dimension control cube 9

10 Sample TI update for Caption = Alias Sample available in technote

11 What about Cloud? CA Cloud + PA Cloud Integration available for existing PA Cloud customers Not currently available for new PA Cloud customers (resolved by EOY) CA On-Prem + PA Cloud Many technical challenges high priority but currently no timelines CA Cloud + PA On-Prem Supported. No special handling required 11

12 CAC + PAC architecture CA Cloud authenticating with on prem LDAP PA Cloud authenticating with BlueID TM1 data tier is configured to Mode 5 authentication which uses CA's CAM component, i.e., SSO running CA reports/dashboards on PA data SSO scenarios User logs into CA (with LDAP credentials), then runs a report with PA data: no prompt for credentials User logs into PA (BlueID credentials), then clicks the "Report" icon (in PA) which opens up CA: no prompt for credentials User logs into PA (BlueID credentials), then pastes the CA URL in another browser tab: you will get prompted for credentials 12

13 OLAP including TM1 and PA in CA Dashboards Why use CA dashboards? Self-service exploration and visualizations One dashboard supports many data sources Known issues with OLAP in CA dashboards: No crosstab summaries (aggregates are available in cells) Zero/null suppression always on (no off switch) Members (elements) do not appear in metadata tree Tip: Select a Hierarchy within a visualization and then Filter 13

14 Optimizing report performance 14

15 Less is Faster When you know which elements have the data you care about, explicitly add those into the report Step-by-step report creation: Add one item at a time and filter that item down as much as possible before proceeding to the next item Read Writing Efficient OLAP Queries on IBM developerworks 15

16 Filtering with a MUN is Fastest MUN = member (element) unique name 16

17 Filter on an attribute value only if you have to Attribute filtering requires a scan through every single element in the set! 17

18 Do not filter with slicers Slicers only remove cell values, never members Detail filters remove members, and they slice Use optional detail filters when adding explicit members to the report is not possible 18

19 Use dimensional functions Efficient Versatile Tips while authoring Tips on developerworks IBM Cognos BI Dimensional Functions 19

20 Use parent members for summaries In most dimensions, every non-leaf member is the rollup of its children As opposed to having summary values computed in real-time, if you use the parent member this value is typically pre-computed TM1 is optimized to return these rollups quickly You can t use this technique if you need a summary that is not a complete set of children, e.g., a TopCount() set 20

21 As a last resort you can create a data set 21

22 Cautions 22

23 PA objects that are currently unsupported in CA Subsets (under development) Views 23

24 Simulating alternate hierarchies by duplicating elements Potential problems in CA Hierarchy navigation functions that depend on member order may not work as expected Users expect NextMember(PYTD. 2014Q4) to return nothing, but it will return 2013Q1 here CA-computed summaries of descendants double-count duplicate members Level expressions may report duplicate elements Level summaries may double-count (while TM1 consolidations are correct) Detail filters may produce incorrect results (double-counting) 2014Q1 will appear twice if a filter selects quarters where Quarter of Year = 1 Solution: Create real alternate hierarchies instead Guidelines if you cannot avoid duplicating elements: Don t combine with level-based reporting Stick to top level sub-trees Avoid navigation across sub-trees

25 When to define Levels in TM1 CA does not require levels in PA cubes Parent-child hierarchies are supported However, levels can make authoring in CA more convenient Defined Levels should only be used when: There is a Balanced Hierarchy Authors need/want to select members by level Attributes are level specific Levels should not be defined when: The hierarchy is Ragged Attributes that span levels are used in reports Levels are defined in the }HierarchiesProperties control cube 25

26 My report unexpectedly come back empty BI suppression relies on skipcheck; being in place in the TM1 Rules file and assumes sufficient Feeders exist See TM1 FEEDERS proven practice on developerworks Underfed cubes will return NULL even when there s data Last resort: to inform the BI server that there are missing feeders, set a cube attribute called IsUnderfed to True (technote ) 26

27 What s New in

28 Text measures in crosstabs 28

29 Planning Analytics connections within Manage Data servers No need for legacy Cognos Administration interface Ability to create data modules on selected cubes 29

30 Thank You