Most of us are familiar with the role that flash cache plays with hybrid storage systems that are a mix of flash and traditional disk. Cache is there as a performance accelerator: storing recent reads, and buffering writes to disk.
But when VSAN 6.0 announced its new all-flash configuration, there was a still a recommendation for cache in addition to flash devices used for capacity. Why is this — aren’t the capacity flash devices fast enough? And why the 10% recommendation?
With all-flash VSAN, cache is used to extend the life of less-wear-endurant (and less expensive!) capacity flash devices. Unlike hybrid configs, cache is not about performance — it’s about economics.
Recently, Cormac Hogan put together an excellent post explaining how cache works differently with all-flash VSAN, and — more importantly — explains the logic behind the 10% usable capacity sizing recommendation.
If you’re into optimized configuration of VSAN — or just want to understand how things work behind the scenes — it’s excellent reading!
http://cormachogan.com/2015/05/19/vsan-6-0-part-10-10-cache-recommendation-for-af-vsan/