April 9, 2020
Today, I want to talk to you about a common problem in storage infrastructure, and that is how to isolate VMs from one another. Sometimes you have databases, a bucket for databases, maybe there's a RAID group, or a LUN, or a datastore that's set aside for database workloads, or maybe just the database log disks, or the databases themselves. Maybe there's a section for virtual desktops. Maybe there's a section for general purpose. Maybe you even have to carve those sections into subsections sometimes to isolate databases from each other.
Well, I've seen environments where people give a single datastore for every VM in an attempt to isolate performance of one VM from another. Because Tintri has gone about this in a different way by creating a file system from the ground up that is able to see a virtual disk and a virtual machine as a primary object, we can actually see what requests are coming in for every VM from a write standpoint and from a read standpoint, and then we can use machine learning and algorithms to isolate and interleave those requests so that now, VMs and workloads are isolated from each other mathematically, rather than physically being separated on different datastores and different LUNs, and that means that no matter what you put into a Tintri array, those algorithms will take care of the isolation and you don't have to spend time planning and managing and deploying distinct infrastructure and distinct disk-backing for every application in your environment. If you want to learn more about this, contact one of our Tintri experts today.