I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

There are a lot of factors that go into this.  That being said, for 10 users you could likely run it on the minimal requirements supported.

Supported Environments
https://www.ibm.com/software/reports/compatibility/clarity-reports/report/html/prereqsForProduct?deliverableId=1413345793248

RAM = Amount of Data
CPU = Data/Concurrent Users

For 10 users on relatively standard models you probably wouldn't need more than 8 Cores and 32GB Ram, however, since you haven't provided details this is only a guess.

For context, we run 1250 users on 32 cores and 512GB RAM although our indiviudal servers don't generally go over about 100GB ram each.

VMs are generally fine for what you have stated.  If you use PAW then Windows Server 2016 is reccomended if you go microsoft.  Linux works for all components however, if you use security mode 5 its easier to setup the gateway on a windows server with IIS since it has the built in capability for handling kerberos/windows authentication

------------------------------
Robby Meyers
------------------------------

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

Thanks Robby. This provides a great starting point and makes sense. We can come up with specs based upon these guidelines.

------------------------------
Shubho Ghosal
------------------------------

Original Message:
Sent: Fri November 01, 2019 02:18 AM
From: Robby Meyers
Subject: Need guidance on optimal server configuration of PA production instance

There are a lot of factors that go into this.  That being said, for 10 users you could likely run it on the minimal requirements supported.

Supported Environments
https://www.ibm.com/software/reports/compatibility/clarity-reports/report/html/prereqsForProduct?deliverableId=1413345793248

RAM = Amount of Data
CPU = Data/Concurrent Users

For 10 users on relatively standard models you probably wouldn't need more than 8 Cores and 32GB Ram, however, since you haven't provided details this is only a guess.

For context, we run 1250 users on 32 cores and 512GB RAM although our indiviudal servers don't generally go over about 100GB ram each.

VMs are generally fine for what you have stated.  If you use PAW then Windows Server 2016 is reccomended if you go microsoft.  Linux works for all components however, if you use security mode 5 its easier to setup the gateway on a windows server with IIS since it has the built in capability for handling kerberos/windows authentication

------------------------------
Robby Meyers

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

The Planning Analytics installation will probably be licensed by number of CPUs. So it is important that you perform a proper sizing before deciding on what your deployment will look like. IBM will be able to help with a questionaire that includes information on number of users, size of your intended data model, global distribution of users, and other factors influencing the size of the cube (RAM) and concurrency of users (CPU). 
With regards to platform, as Robby mentions, review the supported environments page. Planning Analytics can run on many platforms and virtualization technologies. It is important that you also understand the licensing impact of those choices: https://www.ibm.com/software/passportadvantage/pvu_licensing_for_customers.html
For virtual machines there are some documents for TM1 which are still applicable. For instance: https://www.ibm.com/support/pages/guidelines-allocating-memory-virtual-machine-running-ibm-cognos-tm1

------------------------------
Luke Peterson
Business Analytics Advocate
IBM
------------------------------

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

Luke information is very good!

I like to expand on a few things:

• There are many pieces to the cloud as such you must bear that in mind when designing your model
• PAoC now has hubless - https://www.ibm.com/support/pages/using-planning-analytics-microsoft-excel-without-performance-manager-hub
• More and more work, analysis and reporting can be done in PAW - https://www.ibm.com/support/knowledgecenter/en/SSD29G_2.0.0/com.ibm.swg.ba.cognos.tm1_nfg.2.0.0.doc/c_new_features_paw.html
• You also need to look at data loads, i.e. TI processes, MTQ, etc. - https://code.cubewise.com/blog/optimizing-your-tm1-and-planning-analytics-server-for-performance
• There are many areas as such it is important to sit down and mapped out the appropriate strategy when it comes to model design.

*** It is always recommended by IBM for clients to work with either Lab Services or BP as part of reviewing model designs including TM1s.cfg.

------------------------------
paul YOUNG
------------------------------

Original Message:
Sent: Fri November 01, 2019 03:19 AM
From: Luke Peterson
Subject: Need guidance on optimal server configuration of PA production instance

The Planning Analytics installation will probably be licensed by number of CPUs. So it is important that you perform a proper sizing before deciding on what your deployment will look like. IBM will be able to help with a questionaire that includes information on number of users, size of your intended data model, global distribution of users, and other factors influencing the size of the cube (RAM) and concurrency of users (CPU). 
With regards to platform, as Robby mentions, review the supported environments page. Planning Analytics can run on many platforms and virtualization technologies. It is important that you also understand the licensing impact of those choices: https://www.ibm.com/software/passportadvantage/pvu_licensing_for_customers.html
For virtual machines there are some documents for TM1 which are still applicable. For instance: https://www.ibm.com/support/pages/guidelines-allocating-memory-virtual-machine-running-ibm-cognos-tm1

------------------------------
Luke Peterson
Business Analytics Advocate
IBM

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

Thanks, Paul, for the extensive list of inputs. Much appreciated.

------------------------------
Shubho Ghosal
------------------------------

Original Message:
Sent: Fri November 01, 2019 05:21 AM
From: paul YOUNG
Subject: Need guidance on optimal server configuration of PA production instance

Luke information is very good!

I like to expand on a few things:

• There are many pieces to the cloud as such you must bear that in mind when designing your model
• PAoC now has hubless - https://www.ibm.com/support/pages/using-planning-analytics-microsoft-excel-without-performance-manager-hub
• More and more work, analysis and reporting can be done in PAW - https://www.ibm.com/support/knowledgecenter/en/SSD29G_2.0.0/com.ibm.swg.ba.cognos.tm1_nfg.2.0.0.doc/c_new_features_paw.html
• You also need to look at data loads, i.e. TI processes, MTQ, etc. - https://code.cubewise.com/blog/optimizing-your-tm1-and-planning-analytics-server-for-performance
• There are many areas as such it is important to sit down and mapped out the appropriate strategy when it comes to model design.

*** It is always recommended by IBM for clients to work with either Lab Services or BP as part of reviewing model designs including TM1s.cfg.

------------------------------
paul YOUNG

Original Message:
Sent: Fri November 01, 2019 03:19 AM
From: Luke Peterson
Subject: Need guidance on optimal server configuration of PA production instance

The Planning Analytics installation will probably be licensed by number of CPUs. So it is important that you perform a proper sizing before deciding on what your deployment will look like. IBM will be able to help with a questionaire that includes information on number of users, size of your intended data model, global distribution of users, and other factors influencing the size of the cube (RAM) and concurrency of users (CPU). 
With regards to platform, as Robby mentions, review the supported environments page. Planning Analytics can run on many platforms and virtualization technologies. It is important that you also understand the licensing impact of those choices: https://www.ibm.com/software/passportadvantage/pvu_licensing_for_customers.html
For virtual machines there are some documents for TM1 which are still applicable. For instance: https://www.ibm.com/support/pages/guidelines-allocating-memory-virtual-machine-running-ibm-cognos-tm1

------------------------------
Luke Peterson
Business Analytics Advocate
IBM

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson

Thanks, Luke, for the pointers. I will review them for sure. Much appreciated.

------------------------------
Shubho Ghosal
------------------------------

Original Message:
Sent: Fri November 01, 2019 03:19 AM
From: Luke Peterson
Subject: Need guidance on optimal server configuration of PA production instance

The Planning Analytics installation will probably be licensed by number of CPUs. So it is important that you perform a proper sizing before deciding on what your deployment will look like. IBM will be able to help with a questionaire that includes information on number of users, size of your intended data model, global distribution of users, and other factors influencing the size of the cube (RAM) and concurrency of users (CPU). 
With regards to platform, as Robby mentions, review the supported environments page. Planning Analytics can run on many platforms and virtualization technologies. It is important that you also understand the licensing impact of those choices: https://www.ibm.com/software/passportadvantage/pvu_licensing_for_customers.html
For virtual machines there are some documents for TM1 which are still applicable. For instance: https://www.ibm.com/support/pages/guidelines-allocating-memory-virtual-machine-running-ibm-cognos-tm1

------------------------------
Luke Peterson
Business Analytics Advocate
IBM

Original Message:
Sent: Thu October 31, 2019 10:19 AM
From: Shubho Ghosal
Subject: Need guidance on optimal server configuration of PA production instance

I am looking for guidance on what the optimal configuration of a server that would host a production instance of planning analytics would be.

This is a starting configuration and would be released to about 10 users. 

The data set is not large. The interactivity is limited to playing around with some what-if inputs.

The install is prefereed to be on vm hosted on Azure.

Any advice on choice of platform, as well as vm specifications that would be a good starting point (number of CPU threads, memory, etc.) would be much appreciated.

Any links to documentation that is specific to server architecture would also be appreciated, though I am looking myself at this aspect as well.

------------------------------
Shubho Ghosal
------------------------------

#PlanningAnalyticswithWatson