AI-supported, predictive content locations (Super Peers, DPs, Cloud DPs)
Planning your content distribution in ConfigMgr is usually a one-time thing but over time - especially in large environments - optimization of said distribution infrastructure tends to become a "when we have time for it" job while the network topology can change drastically; sometimes without ConfigMgr admins knowing about it. Proper boundaries, branchcache, delivery optimization, pull-DPs etc. already provide great tools for us to optimize this but some questions remain and are currently not answered within ConfigMgr itself:
- which clients could become viable peer cache clients?
- which DPs should act as pull-DPs and from which DP should they ideally pull their content from?
- what benefits would a change in boundary groups/available peer cache clients/introduction of cloud vs classic DP bring?
Answering these questions from within ConfigMgr in management insights, existing dashboards (e.g. the existing client sources dashboard) or - one step further - integrated into Desktop Analytics or combined with AI from the cloud could:
- reduce the manual labor of a ConfigMgr admin to review content topologies
- reduce required skill level to optimize content distribution
- create immediate success stories in form of ROI and business case calculations (we could remove DP X with a super-peer and save X amount of money)
- provide an innovative use of AI to analyze content distribution topology
Inspired by this tweet regarding management insights for super-peer recommendations by Donna Ryan: https://twitter.com/TheNotoriousDRR/status/1199412246817988608