We love the * real-time replication * ease of use when connecting our servers to the storage * level of redundancy inside the box. In terms of the redundancy, with our previous SANs, anytime we wanted to do upgrades or firmware updates, we would have to set aside a maintenance window. But because Tintri has two controllers and everything is redundant, we can do one of them during the day. That means my team will upgrade and fail over from one control to the next, and none of our applications will go down, and our databases don't lose transactions. It's a very steady platform. It's also simple software and integrates well with VMware so we get a lot of information about all of the VMs, how they're performing individually, and about network latency. That's very helpful when you're troubleshooting a slowdown. Prior to using this solution, it was hard to get that kind of information. That was before virtual volumes came out in VMware, which was supposed to be designed to give you more individual information. But Tintri is designed like that out-of-the-box. A user will say, "Hey, this transaction was slow," or "This application seems slow today." A lot of times in the past, when we had an iSCSI or fiber connection, it was hard to know which link in the chain was the problem, or whether it was just the user's perception or the application was busy because it was swallowing a high load at that particular time. Now, we're able to go in and look at what the flash use is by the application and whether there are any network latencies. It becomes helpful because we can eliminate many of those issues and conclude that it was an application-level problem, or the service was busy at that particular moment in time with transactions. We can tell that it's not an underlying hardware or infrastructure issue. I also absolutely love that VMstore enables replication, snapshots, and setting the QoS at the virtual machine level. For QoS, we let the algorithm within the machine itself handle that because our current device is a hybrid T850. It has a set of spinning disks and about five terabytes of flash. You can look at the management console and see how much your applications or VMs are hitting the flash and how many hits are going to disk. The software determines when there are some volumes in your VMs that aren't currently accessed very often and it will put them on the spinning disk. If an indexing program runs and touches those files, or somebody wants to go back and touch files that haven't been touched in a long time, it will read them from disk, which can be slightly slower. But when we will look at the Tintri control panel, our rate of flash hits is almost 99 percent, meaning all of our most important data, the things that people use every day and rely on, are sitting at that flash level and they get the highest level of performance. And we're leveraging more and more of the snapshotting capabilities for disaster and ransomware recovery. It's becoming one of the tiers that we have. We use Zerto for replication backup and to our offsites, and we're now using the snapshots as another tier of backup and restores. But more importantly, we have two separate Tintri units that we use for real-time replication of "protection groups," things that are our most critical systems. They're replicating between the two. The graphical user interface for monitoring performance is sufficient and colorful. You can mouse over the items that you want to see and it will break them out. I don't think you could present that data any better. We've recently upgraded the software and there are some additional improvements, but it has been very user-friendly and very easy to read.
