Cloud is a hot topic these days. Public cloud solutions are available in all shapes and sizes and from all sorts of vendors. However, public cloud isn’t for everyone. A lot of organizations have a heavy investment in their own enterprise cloud. This investment might be in custom applications, or possibly in policies and procedures one may require for compliance. With that in mind, today Tintri announced some new improvements aimed squarely at enterprise cloud. In this post, I am focusing mainly on the improvements made to their scale-out solution.
Anyone who uses Tintri storage knows how simple it is to run. The simplicity is one of its biggest selling features. But don’t believe for a second that it is lacking in features. With simplicity in mind, Tintri is now aiming to bring agility and scale to datacenters, all while maintaining that same ease of use philosophy.
Scaling Tintri VM Stores isn’t quite like a lot of other storage units. Most units can scale by adding new disk shelves to existing controllers. In Tintri’s case, scaling occurs by adding new VM Stores. One of the reasons for this approach is to avoid bottlenecks with controllers. In order to manage all of these arrays, your best bet is to leverage Tintri Global Center. Using TGC you can visualize all of your pooled arrays as one logical unit.
Previously TGC was restricted to managing 32 nodes or 160,000 virtual machines. Going forward there has been an increase to that limit with 64 VM Stores and 480,000 virtual machines. What I find interesting about this solution is that it can be a mix of arrays (e.g. made up of hybrid and all-flash arrays). This bodes well for environments that have a hodge-podge of equipment.
TAKE A LOAD OFF
OK, so an increase in capacity (both nodes and total VMs) is always welcome, but what else is coming down the pipe? A feature that looks to be quite interesting when paired with TGC is a new storage offload engine.
When a lot of us VMware admins hear the term storage offload, we probably think of VAAI (vStorage API for Array Integration). Tintri has supported VAAI for quite a while, and I can’t imagine using an array without it now. Being able to perform a task like deploying a template in seconds is something one can take for granted. But one of the limitations of VAAI is that it is not available across arrays.
As part of their scale-out message, Tintri will bake in their own storage offload engine. A simplified version of what happens when triggering a storage vMotion is:
- The array takes a snapshot (not the hypervisor)
- The array uses its replication engine to move the snapshot to the destination VM Store.
- Of note is that this snapshot will be thin provisioned, deduplicated, and compressed … so optimized.
- When the base snapshot replication is complete, vMotion will transfer the delta
The benefits of this approach largely lie in being able to move less data (because it is optimized) quicker. The process skips using the hypervisor for a large portion of the task. This results in additional overhead reduction as well. What is also handy about this feature is that it will retain existing snapshots and policies.
Along with the scale-out improvements, there is a facelift to their analytics platform. In a growing environment, Tintri’s analytics engine is a key tool. You can easily build models based on profiles that you build. Using these models, you can quickly see what the impact will be on your storage (both capacity and performance). With this new update, you will see compute resources showing up as a metric. These details are directly from the hypervisor and will be stay in Tintri Analytics’ service for 3 years.
Along with the improvements to the Analytics engine, there have been additions to the replication engine. The newest feature rollout includes support for IBM Cloud Object Storage, in addition to the existing S3 capabilities.
Wouldn’t it be great to continually be on the forefront of the latest trends? Absolutely! But, most of us can’t just throw out our existing investment. With these upcoming releases, in addition to an existing automation toolkit (leveraging PowerShell), it is clear that Tintri is keeping its commitment to its customers. In the world of IT, all equipment grows old. Some of it more quickly than others. As a customer and an end-user, it is great to see a vendor assist with getting as much value from an investment as possible.
Disclaimer: Although Tintri is a sponsor of this blog, I was not requested nor required to write this post. Everything written above was written on my own accord.