TM1 Forum

Dear all,

We are doing a financial consolidation project. The TM1 system is used
for data collection and data consolidation. There is no simulation and
sandbox requirement in this project. There are total 60 users. This
system will be busy at month begin from 1st to 10th. The users are
located in different countries. Each user will use this system 3 hours
per day. There are total 250 cubes. 85% of data transformation are going
to use TI. 15% of data transformation are going to use cube rule.

Average number of dimension for one cube is 6. The common dimensions are
company (100 elements), year (10 elements), month (12 elements), account
(1500 elements). The density is 30%.

Does anyone knows any methodology or tool that can help to do TM1 capacity planning?

redamos wrote:Does anyone knows any methodology or tool that can help to do TM1 capacity planning?

I've always found this one of the more difficult aspects of TM1 consulting. Unfortunately I know of no tool or best practice methodology, but that may well be because I generally handball that aspect to the IT department of the client and point them to the proper IBM resources they have at their disposal for support.

When I've been pressed, I err greatly on the size of over-speccing, as I don't want to end up being responsible for a dud hardware purchase.

I would be very interested if there was a tool of this kind, so will be watching this topic closely.

redamos wrote:The density is 30%.

Along with all the other facts and figures you provide, you seem to know this one quite specifically. It makes me wonder how much data you have loaded into your model so far in order to look at real memory consumption patterns? Doing small 'mock-ups' or prototypes can be very useful.

redamos wrote: The TM1 system is used for data collection and data consolidation... The users are located in different countries... There are total 250 cubes... 15% of data transformation are going to use cube rule.

You don't mention what version of TM1 you want to use which can make a difference if you are looking to build a planning application. If your users are spread around the world, do you plan to use local servers and a replication strategy to consolidate the disparate data sets? How do you define 'data transformation' - have you thought about the real processing requirements of the 'transformations'?

redamos wrote:Does anyone knows any methodology or tool that can help to do TM1 capacity planning?

There are so many factors to think about, I'm not sure how a tool is going to beat having human experience on this one. If you're not sure what to do next then it's probably worth paying for someone to come in and help you work out what sort of server(s) you need. IBM ought to be able to provide you with architectural advice and, as Lazarus implies, it is often politically expedient to ask them to do so.