SCCM Design Strategy

How to create SCCM Design strategy

SCCM is very flexible software, has many roles inside that can be installed in small scenarios on one server with all components including SQL server or can be split on many servers with one or more components. Also from redundancy perspective one role can be installed on more servers.

So before start to design your SCCM infrastructure you need to consider the follow details

·         Central Administration Site – A Central Administration Site (CAS) can support up to 25 primary sites. When you use SQL server Enterprise or Datacenter for site database on central administration site, database can support 400.000 clients. In most cases CAS is not a required. You only need a CAS if:

o   You need more than one Primary Site

o   You have more than 100.000 clients

·         Primary Site – A stand-alone primary site support 100.000 clients

o   Each Primary Site Management Point supports 25.000 clients. To reach 100.000 clients you need to have 4 Management Points. One primary site can support 10 Management Points

o   Each Primary Site can support up to 250 Secondary sites

o   Primary site DB, local or remote SQL. Is easier for small environments to install SQL DB local on same server. You can also use a dedicated SQL server or Cluster

·         Secondary Site – Secondary Site needs a permanent connection with Primary Site. If one remote site has less than 500 clients is better to install only a Distribution Point instead of Secondary Site

o   Each Secondary Site supports only one single Management Point and needs to be installed on the Secondary Site Server

o   Maximum number of clients is 5000

·         Distribution Point – Individually each primary site support 250 Distribution Points and each distribution point support up to 4.000 clients. Same for each secondary site.

o   Each primary and secondary site supports up to 2000 additional distribution points configured as pull-distribution points. For example, a single primary site supports 2250 distribution points when 2000 of those distribution points are configured as pull-distribution points.

o   Each Distribution Point supports a combined of total 10.000 packages and applications

·         Boundaries – Best practice is to relate boundaries with IP ranges

·         High Availability – Configuration Manager does not provide a real-time service and you must expect it to operate with some data latency. Therefore, it is unusual for most scenarios that involve a temporary interruption of service to become a critical problem. When you have configured your sites and hierarchies with high availability in mind, downtime can be minimized, autonomy of operations maintained, and a high level of service provided.

In order to assure HA for SCCM environment you can use a SQL cluster for Database and add more instance for some critical roles like Management Point, Distribution Point and SMS provider. For increase redundancy you can do:

o   Use DB on a SQL cluster

o   Create multiple instances of Management Points

o   Create multiple instances of Distribution Points

o   Create multiple instances of State Migration

o   Create multiple instances of Application Catalogue

o   Create multiple instances of Reporting Services

o   Create multiple SMS providers

o   Are not considered mission critical and are without HA options:

§  Asset Intelligence Synchronization Point

§  Endpoint Protection Role

§  Enrolment Point

§  Enrolment Proxy Point

§  Fallback Status Point

§  Out of band Service Point

·         Backup – For backup you need to consider the follow

o   SCCM DB

o   SCCM report DB

o   WSUS DB

o   SCCM packages and folders:

§  SMSPKG

§  SCCMContentLib

§  SMSPKGSIG

§  SMSSIG$

§  SMSPKGE$

§  Network or local apps and packages sources (before to upload in DP’s)