Today, the Central MS VMUG had its Q3 meeting. Since I am a VMUG Co-Leader I was able to help setup this meeting, which was my first. Everyone had a great time networking and learning. The food at Lou’s was great as usual!
The turn out was good, and we had a fantastic presentation by Todd Mace from Datrium!
So what is Datrium?
Datrium is a HCI solution, but is unique when compared to other HCI solutions currently on the market. They offer Datrium turnkey branded compute nodes, or they give you the option to use your own. The core of the Datrium solution is the DVX software, which is the brains behind the solution. Additionally, the DVX Data node is where all of your storage/data is located. Datrium’s DVX System has encryption, dedupe, and always-on compression baked in for truly seamless experience.
How is Datrium different?
Datrium is different in a few ways. First, they allow you to mix and match your own x86 servers (different makes/models with VMware EVC enabled). Secondly, they allow you to BYOF, otherwise known as Bring your own Flash. The compute nodes, like many HCI solutions, require a flash device. This device is used for caching of hot data, which allows for increases performance of applications. This may not seem like a big deal at first, but since your not limited to a specific hardware solution, you have the flexibility to “upgrade” your flash devices within your compute nodes. Flash/SSD/NVMe technologies continue to grow in capacity, but at the same time, are lowering in costs. For example, if you have a 1TB NVMe Flash device, and next year a 2TB NVMe comes out, you just upgrade the device and continue running. With this flexibility, you are able to extend the life of your investments by upgrading, rather than replacing with a forklift.
What is the performance like with Datrium?
Performance with Datrium is based on the number of CPU cycles available for IOPS. Since Datrium offers up to 128 Compute nodes, and up to 10 data nodes, the performance increases as you scale the Compute nodes. How is this possible? Well for one, most typical setups have a storage array. Most storage arrays have up to two controllers. These controllers often cause issues once the workload exceeds the capacity of the controllers. This isn’t so with the Datrium solution. When you increase the compute nodes, performance increases. This not only increases the number of available processors, which are used for I/O, but also the flash cache available. The performance is directly tied to the amount of available CPU cycles you have left over after your workloads have been taken into consideration. In terms of the performance, they are two modes: Fast and Insane. The fast mode can use up to 20% of the available CPU cycles, and insane mode can use up to 40% of the available CPU cycles. I love the Insane hats! Get it?!?
What protection options does Datrium provide?
Datrium provides a built in snapshot feature, which can be used as a backup solution. These are very efficient snapshots, which are taken without the typical VM “stun” associated with VMware snapshots. If you require application consistency, then all you need to do is install Datrium’s VSS provider within the guest OS. This will allow for application consistency when snapshots are taken. In terms of data offsite, you can ensure your primary data is offsite in a cloud DVX. Currently Datrium has an AWS solution that allows for this.
In closing, the Datrium presentation at our Q3 VMUG Meeting was a success!! Thanks Todd. If you have any information, be sure to reach out to Todd @mctodd on twitter. More information about Datrium can be found at their website.