The scalability of Tintri is where it could possibly be expensive. 

When we had LeftHand SAN, each storage node was like a mini server so to grow your storage you had to buy a very expensive device to put in there. That would increase the spindles, computing power, and throughput. You got everything, but if you just needed storage, you couldn't do that.

With Tintri, you're doing the same thing. Now, it does allow you to add a unit and then cluster them together to increase the volume size of your scale out. If you are just looking for storage, and you're having to scale out by buying a whole new Tintri node, you're getting additional compute and everything else. I'm not sure that it's the most cost-effective way of doing it. It might be better, instead of scaling it out in their traditional sense, to just add a node, carve it up, and have it be another target for VMware. 

However, in a lot of the newer storage arrays from Tintri, you have a lot more flash space. If you're only buying a unit at a certain size, there are empty slots for you to add to it. That aspect is different with the flash version versus our T850, which was fully provisioned. It's a box of spinning disks and flash. You don't grow it that way. But you can grow the newer units by adding more flash disks.

The scalability internally through VMstore is quick and easy. Scaling up the physical boxes is also dead easy. You buy another box, plug it in, connect it, and away it goes in 30 minutes.

The scalability is good. 

One thing that is a little unique about Tintri is that, while you can present a large pool of storage, there are some idiosyncrasies around how that is seen by VMware. However, Tintri Global Center makes it appear as a single storage volume because it moves workloads predictively.

It's very scalable.

We have been easily able to scale from 100 to 500+ VMs using Tintri VMstore; we have had no challenges scaling up while using Tintri VMspot

This is our main love of these units -- connect to VMware and Tintri monitor and manage your performance.

Although now we have SSD drives, during the time that I was using it, it was mostly physical drives. What I feel would be nice, in terms of a wishlist, is scalability. Rather than replacing the whole appliance, I would like to be able to just add another unit and scale it like that. I know there are some limitations to things like that, but scalability, as opposed to having to buy a new appliance for higher capacity, would be good.

We have not encountered any scalability issues.

We've had no issues with scalability.

When Tintri gets to either the maximum performance reserves or available disk space, the only upgrade path is to purchase a new unit.

No issues.

We currently only have one device, so have not scaled it.

We've been able to scale it without issue.

We've had no issues with scalability.

None

Scalability is just an add-more proposition. It is easy and the units come essentially in small, medium, and large.

At the moment, it is half-filled with SSDs; it has have 25 TB in SSD capacity (12.6 TBof it is usable) and after dedupe/compression, 80 TB is usable.

We have not had any issues with this. Currently, we only have one VMstore.

There were no issues with scalability. While we didn’t have the need to add shelves, that should be a straightforward process

Scaling is via a loosely-coupled "cluster" of independent VM store appliances.

No.

Scalability is not really an issue. When you run out of space, you just purchase another unit and create a new share.

Moving on to a synched and automated mirror between two data centers was not possible. We have been hearing for a month that this feature will become available. However, in the end, it was too late for us.

For now we don't have scalability issues, but It would be nice to add an external bay for additional storage space. If you need more space, you need to buy another full VMStore.

Not until now. Because we don't have a full flash array but a hybrid we expected to not have a 100% flash hit ratio but to this day we never got below 100% hit ratio of Flash.

Scaling requires the purchase of a new VMstore.

We currently have about 240TB of data and add about 60TB a year. Ideally, we prefer to have a single array to manage for our primary workload. Our VDI environment is stored separately and Tintri more than fits the need of this environment.

We've been able to scale without any problems.

We encountered no issues with scalability.

Not yet.

We've had no issues with scalability.

No. We're onto our 4th unit.

Scaling is very easy to do, as they effectively are standalone units. You can get a Global Center OVA which monitors and manages up to 32 devices at once. We have this in-house and use it frequently.

It scales as promised. You run out of resources, you buy another one. We have not hit the limit on any of our boxes yet.

No issues.

We haven't had scalability issues.

We haven't had to scale it that much yet, but we haven't had any issues.

It's scalability is great. Because it’s node-based, if you need more space, you just add a node.

Well, this is one of the most important points: 1 rack for almost 1 PB!

We are a medium-sized enterprise. We have around 200 users. 

We filled up the performance quota before filling up the space.