First published on TECHNET on Dec 20, 2018
As 2018 wraps up we are starting to set our sights on what the future of our System Center Configuration Manager (ConfigMgr/SCCM) infrastructure could bring. Our move to Azure was focused on learning and proving what could be done. Now our thoughts turn toward cost reduction. Here are a few things on our wish list for the coming year:
Scale back virtual machine sizes
– We built machines with a bit of caution in our planning. Now that our migration is complete and we have a few months of real running numbers we want to scale down our VM sizes in a few spots and save some money…, A nice ability that Azure gives us.
Cloud Management Gateway redeployment
– We were early adopters of CMG, which were built on Azure Classic VMs. We also had separate cloud distribution points (DP). With the role combination in the new ARM based CMG we are looking to replace two web services with one and simplify the management of it all as well.
Dynamic site systems
– Azure has the nice ability to do auto scale-up and down of web services. We want to investigate doing this for management points (MP), DP and such to see if there is a way we can dynamically scale up and down our site systems based on load. This is not going to be a simple task but we hope to get some help from the developers to make it more feasible and then flush out the rest of the scenario with some custom scripts.
– When we stood up our initial CMG we had reasons to have them map back to specific, dedicated, MP. Through bug fixes and design change requests we no longer have that need so we want to remove that extra MP and re-point CMG to use the existing MP architecture.
Remove internet load balancers (ILB)
– Due to legacy design and needs we have/had windows software update servers (WSUS) behind an HLB/ILB. As we revisit those needs we hope to eliminate the need and be able to get back to the standard WSUS list functionality built-in to the ConfigMgr product.
– Our client support needs are shifting. With more devices doing azure active directory (AAD) enrollment and Co-Management loads shifting to Intune it is worth investigation and discussion with the ConfigMgr development team about the scale numbers for a site and what we can possibly do to combine our primaries and reduce costs in that method as well.
Remove unnecessary site systems
– With our peer cache success, the number of DPs can reduce (perhaps to only 1 per site, with dynamic scale up when needed). New MP functionality lets us start the removal of app catalogs. MP scale numbers can be re-assessed to possible drop one or two MP per site.
Co-Management workload transition
- We are going to continue to test and transition more workloads to Intune including things such things as windows update, endpoint protection, and device configuration.
As we know, we don’t always get all our presents under the tree. Hopefully we will get most of them. Happy holidays to you all!