3.ibm cognos-tm1-9.5.2-why-upgrade

© 2011 IBM Corporation
Information Management
TM1 9.5.2 – Why Upgrade?
Brian Simpson
Product Manager – Cognos TM1

2
Important Disclaimer
 THE INFORMATION CONTAINED IN THIS PRESENTATION IS PROVIDED FOR INFORMATIONAL
PURPOSES ONLY.
 WHILE EFFORTS WERE MADE TO VERIFY THE COMPLETENESS AND ACCURACY OF THE
INFORMATION CONTAINED IN THIS PRESENTATION, IT IS PROVIDED “AS IS”, WITHOUT WARRANTY OF
ANY KIND, EXPRESS OR IMPLIED.
 IN ADDITION, THIS INFORMATION IS BASED ON IBM’S CURRENT PRODUCT PLANS AND STRATEGY,
WHICH ARE SUBJECT TO CHANGE BY IBM WITHOUT NOTICE.
 IBM SHALL NOT BE RESPONSIBLE FOR ANY DAMAGES ARISING OUT OF THE USE OF, OR OTHERWISE
RELATED TO, THIS PRESENTATION OR ANY OTHER DOCUMENTATION.
 NOTHING CONTAINED IN THIS PRESENTATION IS INTENDED TO, OR SHALL HAVE THE EFFECT OF:
– CREATING ANY WARRANTY OR REPRESENTATION FROM IBM (OR ITS AFFILIATES OR ITS OR THEIR SUPPLIERS
AND/OR LICENSORS); OR
– ALTERING THE TERMS AND CONDITIONS OF THE APPLICABLE LICENSE AGREEMENT GOVERNING THE USE OF IBM
SOFTWARE.

3
TM1 9.5.2 Enhancements
Read / Write Scalability & Performance
• Parallel Interaction
• Data Reservation
BI Integration
• BI Report Performance
• TM1 iWidgets
Contributor Usability
• Multi-Node Edit
• Text Wrap
• Spreading Shortcuts
Conformance
• Red Hat Linux*
• Excel 2010
• BI Pkg. Con. / C10
Language
• New Languages
Rules & TI
• Min, Max, Avg
• Count, DCount
TM1Top
• Sandbox
• Job Queue

4
Parallel Interaction
 During periods of read/write contention,
writer performance is improved
 Performance is steady as users scale
 Analysis of updates in real-time
 Concurrent data spreading
 Eliminates Writer Partitioning Best Practice
 Intra-Day Data Imports do not inhibit
writers
 Intra-Day TI Processes do not inhibit
writers
 Faster parallel data loads to same cube
 Scale leverages server core capacity
 Job Queue multi-threads processeing
Goal
• Greater scale for read/write
environments
• Faster performance for writers
• Simultaneous Activities (data
maintenance writes / TI & reads)
• Maintain reader performance
• Greater CPU utilization
Benefits
Notes:
• PI provides benefits when contention exists
• PI is not targeted at reader performance
• Meta Data maintenance still blocks
• Memory consumption will rise ~ 10 – 30%
• Watch for core saturation as you scale

5
Demonstrating 9.5.1 Locking Behavior – a brief review
Cubes linked by rules will „lock‟ together - A read to
Cube A will lock a write to Cube A & B
Rules create a
dependency
between Cube A
& Cube B
Cube A
Cube B
Blocked writers will „stack up‟ behind concurrent
readers and wait for the reads to flush out.
Readers are free to process together, so multiple
reads can occur in parallel to all cubes
When the cube is „free‟, the writers process serially
(they block each other)
ReadersWriters

6
Without Concurrency, 9.5.1 performance will be on par with 9.5.2 PI
This is a TM1 9.5.1 example demonstrating behavior
without concurrency conflicts
These readers and writers are not conflicting with each
other, therefore there is no waiting occurring due to
object locking
9.5.2 with PI will not improve performance of this
scenario because there is no concurrency conflict

7
9.5.2 PI Enhances Writer Performance during High Concurrency
– No Waiting!
TM1 9.5.2 PI removes object locking that occurs as
a result of data reads or writes
Writers are no longer blocked by readers (or
other writers) – they process without delay*
9.5.2 PI performance improvement vs. 9.5.1 is noticed
when in 9.5.1 scenarios demonstrating contention - and
when there are sufficient Server Cores to handle the
greater level of concurrent transactions
* 9.5.2 still has object „locking‟ scenarios
caused by Meta Data updates (including
Dynamic Subsets, element updates, etc..)
….. We‟ll fix that in another release 

8
Existing TM1 Read Concurrency
Cube A
Read 1 Cube A
Read 2 Cube A
Read 3 Cube A
Processing Time
Completion

9
Existing TM1 Read/Write Concurrency
Cube A
Read 1 Cube A
Read 2 Cube A
Read 3 Cube A
Write 1 Cube A
Write 2 Cube A
Processing Time
Completion
Wait Time

10
Read/Write Concurrency with Parallel Interaction
Cube A
Read 1 Cube A
Read 2 Cube A
Read 3 Cube A
Write 1 Cube A
Write 2 Cube A
 Read 2 will not include data impact from Write 1 (because it
begins before Write 1 completes)
 Read 3 will include data impact from Write 1 (because it
begins after Write 1 completes), but not Write 2

11
How does PI work?
Cube A
Read 1 Cube A
Read 2 Cube A
Write 1 Cube A
Read 3 Cube A
Write 2 Cube A
tt+1t+2
Time line
 TM1 previously manage a single
Data Tree to Access Cube Data
 Reads could share a Tree, but Writes
had to wait for Exclusive Access to
update the Tree
 Parallel Interaction creates a new
Data Tree Access Point „version‟ for
each Write, allowing concurrent
Reads & Writes
 Reads access the latest Data Tree
Access Point to get the most recent
updates
Access „trees‟ to cube
data are versioned,
then updated

12
Starwood Read/Write Concurrency Test
25.4 37.3
189.7136.7
499.8
2003.4
0
500
1000
1500
2000
250 Users 500 Users 1000 Users
TM1 Release
AVGAggregateResponse
Time
9.5.2 9.5.1
Read / Write Concurrency Improvements 9.5.1 -> 9.5.2 PI
Windows 2003 / 8 Core
Load Runner / Cube Views
2 Hour Test
9.5.1 -> 9.5.2 Optimizations C-
Lock, Data Spreading, and Auto
Recalc (very applicable to this test)
Server‟s 8 cores
maxed out @
1000 usersPI runs
93% Faster
PI runs
81% Faster
PI runs
90% Faster

13
0
10
20
30
40
50
60
70
80
90
100
00:00
00:04
00:08
00:12
00:17
00:21
00:25
00:29
00:34
00:38
00:42
00:46
00:51
00:55
00:59
01:04
01:08
01:12
01:16
01:21
01:25
01:29
01:33
01:38
01:42
01:46
01:50
01:55
01:59
02:03
Duration of test (hh:mm)
%ofTotalCPU
0
10
20
30
40
50
60
70
80
90
100
:00
:17
:34
:51
:08
:25
:42
:59
:16
:33
:50
:07
:24
:41
:58
:16
:33
:50
:07
%ofTotalCPU
0
10
20
30
40
50
60
70
80
90
100
0:00
0:08
0:17
0:25
0:34
0:42
0:51
0:59
1:08
1:16
1:25
1:33
1:42
1:50
1:59
2:08
2:16
2:25
2:33
2:42
2:50
2:59
Duration of test (hh:mm)
%ofTotalCPU
1000 User R/W
500 User R/W
250 User R/W9.5.2
9.5.1
9.5.2
9.5.1
9.5.2
9.5.1
Lower Volume
not taxing 8
core server
500 User test
reaches CPU
capacity
1000 Users
saturates server -
delays due to
CPU constraints

14
Parallel Interaction Notes
 No benefit / possible negative impact to Read performance
– Reads now compete with Writes for system resources
– Availability to cache is lessened (due to more frequent write invalidation)
– Separate Reader environments still a „Best Practice‟
– Read Only environments operate with PI disabled
 CPU utilization will be greater, raising importance of server core capacity
– PI performance benefits are reduced when server CPU power is fully utilized
– Increase cores in alignment with concurrent read/write scale
 Dimension Updates still wait for Reads and block other Dimension Updates
– Isolate Meta Data Updates from Data Updates in TI Processes
 Other activities remain subject to „blocking‟
– Save Data All
– Views with Dynamic Subsets
– Public Views with UDCs (including Subsets in a Subset)
– 1st Time View activity following Cube Loading (Cube Dependencies)
 Best Practice - Separate Meta Data Loads from Data Loads to lessen „lock duration‟
– Run TI command line

15
BI Integration
 TM1 9.5.2 Conforms with C8.4 and
Cognos 10
 BI Reporting on TM1 is faster
 BI Server memory not a bottleneck
 Zero Suppression reporting against large,
sparse TM1 Databases having big
dimensions is much faster
 Top Count, Measure filtering, and
Attribute filtering reports are much faster
 Reports indirectly referencing members via
levels or „children of consolidation‟ are
faster
 Drag „n Drop TM1 iWidgets to Business
Insight dashboards
 TM1 iWidgets adopt BUX toolbar
Goal
• Cognos 10 Conformance
• Improve BI Report Performance
against TM1 Data Sources
• Leverage TM1 for suppression
and filter reports
• Optimize MDX generated by BI
• Reduce network data transfers
• Polish TM1 iWidgets for Business
Insight
Benefits
Notes:
• Supports Cognos 10 „Dynamic Query‟ mode
• BI Package Support for DMR / SAP BW

16
Single Portal with Zero-footprint Web Interface
RDBMS DW OLAP Data
TM1 OLAP
PowerPlay
SAP BW
ESSBASE
MS Analysis
Services
Heterogeneous Flat Files
Common Metadata, Security, Integration Services, Query Engine, Automation, Process Mgt
Cognos BI with TM1
Google
Search
Cognos
Mobile
Cognos
Office
BI Analysis
For Excel
Reporting , Dashboarding, Ad-hoc Query, Analysis, Scorecarding, Event Management
TM1 Web, Executive Viewer

17
Cognos 10 Business Insight Dashboard with TM1 iWidget

18

19
Excel Planning Templates What-If Analysis
Web Reports BI Dashboards & Scorecards

20
Cognos 10.1 FP1 on TM1 9.5.2: Moderate Sized TM1 Datasource Reporting
Only the first report execution time is
measured as this more closely matches a
“real life” scenario of a user running a report
once in a given user session.

21
Contributor Usability
 Planners having budget responsibility for
multiple nodes (cost centers, projects, etc.)
can update some/all nodes in a single
Contributor Session
– Perform update while comparing nodes
– Data Spread across nodes
 Planners can submit all nodes of a single
consolidation in one action
 Approvers can review / update a
submission of multiple nodes in one
session
 Workflow page presents a single hierarchy
for dual role Approvers/Contributors
 Text measures used for commentary can
have fixed column widths and height
 Consolidation data spreading when all
children are null / zero
Goal
• Extend Contributor 9.5.x
• Multi-Node Editing
• Enhance Contributor Workflow
page layout and content
• Provide pre-defined column width
and text wrapping configuration
• Enhance data spreading
intuitiveness
Benefits
Notes:
• Next TM1 release has many Contributor
functional and scale enhancements

22
Single Workflow Hierarchy
Single Hierarchy
for Approvers /
Contributors

23
Workflow Page Re-Design
Permissions appear
in grid view
Ownership node
identifies single or
multi-node package

24
Text Column Wrapping (for Contributor UI and Web Cube Views)
Edit
Web.config configuration for
all text columns (character
width and count)
Web.config setting

25
Consolidated Cell Spreading: Auto Proportional / Spread to All Leaves
Spread lump sum to
all leaves : spread
even if all leaves are
null/zero
Enter 52k in Total Year Equipment
ProportionSpreadToZeroCells=T
ProportionSpreadToZeroCells=T
Tm1s.cfg

26
Conformance
 Red Hat Enterprise Linux 5.3 (Introduced with 9.5.1 Silent Refresh Pack)
 Windows 2008 for TM1 Workflow (Introduced with 9.5.1 Silent Refresh Pack)
 Excel 2010 Client and Web Server

27
Languages
 Traditional Chinese
 Korean
 Brazilian Portuguese
 Russian
 Polish
----
 More CEEMA languages coming in 2011

28
New Rule and TI Functions for Statistical Analysis
• Min, Max, Avg, Count,
Distinct Count
• Min, Avg, Count can
include or exclude zeros
• Count is a multi-
dimensional count
• Distinct Count identifies a
single dimension for
counting distinct
occurrences of a value

29
TM1Top Enhancements
 Expand Capabilities of TM1Top to :
– Monitor Sandbox Activity
– Monitor Job Queue Activity
 Enabled by new TM1Top.ini commands
 View Sandbox Memory impact / # of
Sandboxes
 View Job Queue activity, kill Jobs

30

Questions & Comments

3.ibm cognos-tm1-9.5.2-why-upgrade

More Related Content

What's hot (15)

Viewers also liked (16)

Similar to 3.ibm cognos-tm1-9.5.2-why-upgrade (20)

3.ibm cognos-tm1-9.5.2-why-upgrade