Determining the size for both RapidMiner Studio and Server

RG2000RG2000 Member Posts: 4 Learner I
edited June 2019 in Help
I am trying to size the number and size of licenses for both Studio and Server. Here are a few data points

a) I expect the usage to be well-below 'Big Data'. I suspect we will be looking at records under 100,000 records (and if we are lucky to have long history, around 500,000). 
b) The data will get used creatively for multiple use cases- so I do expect a large number of models (10-20) built and deployed- with typical usage being under 10,000/ day and reasonably distributed through the business day. I am struggling to judge how complex the models could be (likely to be simple)
c) I reckon we would have 3-5 business analysts reviewing the models and probably 1-2 data scientists working at a code level

Given these, I had the following questions
1) would a combination of small & medium studio licenses suffice- with the expectation that I can upgrade from small to medium if need arises?
2) On the server, would the basic server be sufficient. Latency is not critical- but availability is (would I need to plan for HA?)
3) Is there limitation on the number of models that can be created/ deployed on the server


Tagged:

Answers

  • sgenzersgenzer Administrator, Moderator, Employee, RapidMiner Certified Analyst, Community Manager, Member, University Professor, PM Moderator Posts: 2,959 Community Manager
    hi @RG2000 I would strongly recommend reaching out to our sales team or partners. They will best equipped to recommend the right setup for your use case. 

    Scott

  • SGolbertSGolbert RapidMiner Certified Analyst, Member Posts: 344 Unicorn

    from your description it seems that Medium licenses for both Studio and Server would suffice. When the need arises, upgrading is very straightforward.

    The number of models that can be deployed is limited by the memory and number of processors in the server. What's really important is how many will be accessed at the same time. If the access is relatively sparse, you can have hundreds of models deployed. The training frequency of the models could be a problem too.

    If you are located in Germany, we can help you (I work for one of the partners).

    Regards,
    Sebastian



Sign In or Register to comment.