vSphere 6, vSAN 6 & Other Key Announcements from VMware PEX
Well, there’s nothing like coming back to the beautiful 4 ft. of New England snow after having been in the temperate climate of the Bay Area for the past week. Might be time to consider becoming a snowbird! In any case, there was a lot of news coming out of the VMware Partner Exchange (PEX) event over the course of the past week. The 3 major announcements were vSphere 6.0, vSAN 6.0, and the VMware/Google partnership. There was also some interesting news from EMC in relation to their highly anticipated launch into the hyper-converged market and the announcement of VSPEX Blue. Today, I’ll cover the highlights of these announcements starting with vSphere 6.0.
vSphere 6.0 represents one of the, if not the, biggest launches in the history of VMware. The core themes of vSphere 6 are scale and elasticity. I won’t go through every new bell and whistle in this post but will focus on the highlights which include increased scale, cloud readiness and elasticity, storage, and high availability improvements. First, on the scaling front, basically, everything has doubled from vSphere 5.5: 64 hosts per cluster rather than 32, 12TB of RAM per host, 480 CPUs per host, etc. When it comes to individual VMs, the same holds true with support for 128 vCPUs and 4TB RAM per VM. I would love to see a system that runs VMs of that scale!
In the cloud readiness/elasticity arena, we now have more truly federated vCenters with shared catalogs, templates, etc. between them. WAY better than the simple Linked Mode of the past. We also finally have the long-awaited long-distance vMotion capability supporting up to 100ms of latency and breaking the old layer 2 network boundaries. However, beware of the large pipes required to really make it sing! Perhaps one of the most interesting new features is Instant Clone, which allows instantaneously cloning a running VM in memory. This is going to be a great leap forward for developers, virtual desktop environments, or anywhere else where fast cloning can be utilized.
On the storage front, we saw the official introduction of Virtual Volumes (vVOL) into vSphere. Essentially, vVOL enables storage management at the VM rather than the LUN level which can greatly simplify management. This has been talked about for several years but is now finally a reality and we should see the majority of storage vendors offering supporting solutions very soon. We also saw that vSphere Data Protection Advanced (vDPA) is now just rolled into the vSphere product rather than requiring additional licensing. If you’re an EMC Avamar customer, this is great news as you’ll be able to integrate and replicate your vDPA backups to your physical Avamar appliances. If you’ve been looking at vSphere replication, there are some great enhancements there as well, including network compression and fast full sync. In the HA area, we’ve long-awaited multi-vCPU(up to 4) support for Fault Tolerance. I believe we’ll see some actual use of this cool new feature now that it can protect higher-end VMs.
vSAN 6.0 was rolled out as part of the vSphere 6.0 announcement. As you probably know, vSAN is the idea of taking local server storage across multiple hosts and clustering it together to create a pool of primary storage capacity without the requirement of a traditional external shared storage architecture. vSAN 1.0 was released a little more than a year ago and is the underlying foundation of the EVO:Rail hyper-converged solutions on the market today. The problem was, while it did work, vSAN 1.0 was missing several capabilities required to really bring it into the production primary storage conversation. Many of those missing links are now filled in with vSAN 6.0.
vSAN 6.0 now supports an ‘all flash’ configuration allowing persistent data to be stored on the flash drives, whereas in 1.0 the flash was used only for caching. We also have a new file system format with vSAN 6.0, providing much more efficient snapshots and increased overall performance. Support for VMDKs up to 62TB and up to 64 vSAN nodes in a cluster bring it online with the new vSphere 6 max configs.
On the HA front, with vSAN 6.0, you can now have fault domains. This basically gives you the ability to recover from a full rack failure, as well as a host failure (assuming you have a good number of hosts in your cluster). Finally, there is greater visibility from a health and troubleshooting perspective built into vSAN 6.0, which should allow it to find its way into more production environments.
The final big announcement at the event was the partnership with Google to provide some of the Google cloud services within the vCloud Air platform. My colleague Tim Cook will be posting a separate segment covering the details of that partnership.
So, when can you download the bits and get all of this goodness in your own environment? Well, I don’t have a hard date, but my guess is we’ll see the GA code released sometime before the end of March. As always, feel free to reach out if you would like more information.