Difference between using Machine Learning Services in SQL vs having a Machine Learning Server RRS feed

  • Question

  • Hi All,

    I am trying to understand the key difference between using Machine Learning Services in SQL vs having a Machine Learning Server.

    Can anyone tell me what is the difference and what scenarios are they used for and point me to any material.



    Thursday, March 29, 2018 2:45 PM

All replies

  • Hi Hermanth,

    As such there is not much of a difference, they both do the same thing (more or less). When I look at it, for me it comes down to where you have your data: if you have your data in SQL Server, then it makes sense to use SQL Server Machine Learning Services. On the other hand; if your data is "all over the place", then perhaps Machine Learning Server fits better. | @nielsberglund

    • Edited by nielsbMVP Saturday, March 31, 2018 6:21 AM
    Saturday, March 31, 2018 6:20 AM
  • I recently stumbled upon similar dilemma, let's say you have your own hosted (not Azure cloud) APS/PDW (which doesn't support/have ML or R integrated) system/appliance instead of single SQL Server (perhaps with only failover or balancer). So unless you have separate ETL/SSAS/SSRS server(s) which would require additional SQL Server License (then you can integrate R/ML into one of those to avoid extra Licensing costs) you would need a place/machine(s) to host ML services - last point also applies if you want to distribute workloads between machines in a farm (also considering other factors: independence on upgrades/administration, failures, service restarts, high workload time windows, etc.)

    Plus from my understanding ML Server is better from scalability perspective - so you can distribute calculation between multiple compute nodes on different servers.

    Saturday, March 31, 2018 8:58 AM
  • Dear Hemanth,

    SQL Server ML Services is an In-database feature fully integrated with the SQL database engine whereas the ML Server is an standalone installation which you can install it without the SQL Server.

    If you are looking for a enterprise node architecture setup, the ML Server can be used to setup the machine learning server and connect to the SQL Server which is in the diff system.

    Refer below for more information on this

    • Proposed as answer by renebouw Saturday, April 14, 2018 7:00 PM
    Thursday, April 5, 2018 5:41 PM