locked
SCCM 2012 Sizing RRS feed

  • Question

  • Hi Guys,

    Need some suggestions here. We have about 12000 clients - 4k in US, 4k in Europe and 4k in China. Each region has a DataCenter. There are 30 branch Offices in each region - (30 US, 30 Europe and 30 in China). All (Most) offices have atleast more than 100 clients.

    Our SCCM 2012 Plan would be to have a CAS and Primary Site in Europe (Head Office) and 1 primary at USA DC, 1 Primary at China DC.

    Each Branch Office to have Secondary Site installed on a virtual machine.

    Does this make sense?

    Wednesday, June 20, 2012 5:41 AM

Answers

  • 1) You only need a CAS if you scale beyond 100,000 clients

    2) In ConfigMgr SP1 you will be allowed to introduce a new CAS to an existing primary site.


    Kent Agerlund | My blogs: blog.coretech.dk/kea and SCUG.dk/ | Twitter: @Agerlund | Linkedin: Kent Agerlund

    • Marked as answer by RookieSCCM Wednesday, June 20, 2012 9:07 AM
    Wednesday, June 20, 2012 6:32 AM
  • Why making things complicated? ;-) Just follow Kent's suggestion to keep it simple. You could of course set up a CAS, multiple primaries and lots of secondaries if you want to waste money and resources.


    Torsten Meringer | http://www.mssccmfaq.de

    • Marked as answer by RookieSCCM Wednesday, June 20, 2012 9:07 AM
    Wednesday, June 20, 2012 8:42 AM

All replies

  • No, it does not make sense :-) Configure a primary site in the main data center. A secondary site in each of the larger remote officies with more than 500 clients and a remote DP in the other remote officies.

    Kent Agerlund | My blogs: blog.coretech.dk/kea and SCUG.dk/ | Twitter: @Agerlund | Linkedin: Kent Agerlund

    Wednesday, June 20, 2012 6:01 AM
  • Thanks for your reply Kent. Basically, the reason to choose SS in every site is.

    There is good and fast network communication between all DataCenters. However, the Branch offices only have good connection with their DC's in their own geographical region.

    For example - in US we have a DC in Houston. The branch offices in Austin, Miami, New York etc have reliable network connectivity with the server in the DataCenter in US only. They have limited connectivity to other DC's in Europe and China.

    If we set the only 1 PS Site in Europe, and DP's in most of the BO Offices - then BO offices from China and US will have network issue connecting to the office in Europe.

    AFAIK - Communication between a SS and a Primary is always compressed opposed to a having a DP in remote location <- In this case clients will send uncompressed data to the MP in DC (Europe)

    Wednesday, June 20, 2012 6:09 AM
  • I'm OK with choosing secondary sites. But there is no reason for you to have a CAS and multiple primary sites. With secondary sites you can also use content routing. that enables you to control that traffic to a secondary site in Austin is routed thru another secondary site Houston.

    Kent Agerlund | My blogs: blog.coretech.dk/kea and SCUG.dk/ | Twitter: @Agerlund | Linkedin: Kent Agerlund

    Wednesday, June 20, 2012 6:21 AM
  • Understood, now I will go with only 1 Primary Site in Europe and Secondary Site every where else.

    The thing about going with a stand alone primary site is that if we have to introduce CAS in the future, we will not be able to (atleast in this hierarchy) without a rebuild kind of thing. Am I correct?

    Wednesday, June 20, 2012 6:27 AM
  • 1) You only need a CAS if you scale beyond 100,000 clients

    2) In ConfigMgr SP1 you will be allowed to introduce a new CAS to an existing primary site.


    Kent Agerlund | My blogs: blog.coretech.dk/kea and SCUG.dk/ | Twitter: @Agerlund | Linkedin: Kent Agerlund

    • Marked as answer by RookieSCCM Wednesday, June 20, 2012 9:07 AM
    Wednesday, June 20, 2012 6:32 AM
  • Kent,

    Thanks for all the help so far. If we have dedicated servers available in all Branch Offices for a ESD Solution and we are implementing SCCM 2012.

    Would it not make sense to make use of it and have Secondary Sites on all BO offices - except offices which are very small?

    What are the examples or implications of this - if you think this is over sizing?

    Wednesday, June 20, 2012 7:34 AM
  • Why making things complicated? ;-) Just follow Kent's suggestion to keep it simple. You could of course set up a CAS, multiple primaries and lots of secondaries if you want to waste money and resources.


    Torsten Meringer | http://www.mssccmfaq.de

    • Marked as answer by RookieSCCM Wednesday, June 20, 2012 9:07 AM
    Wednesday, June 20, 2012 8:42 AM
  • I concur with Torsten and Kent.

    In general, you are using the opposite process than what you should use to design your infrastructure. You should start with the simplest design and then find reasons why it will not work and then add to the hierarchy to address those reasons. Starting complex and trying to simplify is backwards and will end up costing you money.


    Jason | http://blog.configmgrftw.com | Twitter @JasonSandys

    Wednesday, June 20, 2012 2:01 PM