Every year, VMware has new product and strategy announcements that steal the show at VMworld, so vendors tend to make their own announcements in the two weeks leading up to the conference. This year is no different, with nearly all vendors in the virtualization space announcing and readying their latest wares for the 25,000+ people who attend. Let’s look at the bigger themes in play this year and think about how to differentiate the competition at VMworld 2014.
Network Virtualization at VMworld
Network virtualization in the form of VMware NSX was unveiled last year, but like all 1.0 products, it had a ton of rough edges and missing features, and annoying details that made it hard for customers to justify switching. As time has passed, those things will be fixed and compelling use cases will appear. Data center networking is ripe for silo-busting and serious change: look new and particularly interesting integrations from traditional networking vendors, especially in the security space. Be wary of vendors who just port their traditional hardware approach to NSX, though. In many of those cases, NSX can do what you need natively, and you’re just spending money to avoid retraining or reintegrating.
Hyperconvergence at VMworld
Converged infrastructure has been interesting to many organizations, but it has never really achieved the operational efficiencies that it promised. Hyperconverged infrastructures, which take commodity servers filled with CPU, RAM, disk, and network and build a self-sufficient cluster from those raw materials, promise tremendous value on both the CapEx and OpEx fronts. Talk is cheap, though, and just because a company calls their product “hyperconverged” doesn’t mean it is.
A good hyperconverged solution can do everything a traditional virtualized data center can do, except it takes much less money and labor to do it. Hyperconverged solutions are incredibly complex under the hood, while feigning simplicity on the front end, so it’s important that you ask good questions to find the seams and gaps in the technology. Good questions are those around patching, hardware and software support, the time it takes for a vendor to support a new version of vSphere, and storage technologies like deduplication, tiering, and replication. Don’t forget that with most hyperconverged solutions, you’ll need to bring your own networking to the table, too. So if you’re into silo-busting in your own organization, keep that in mind.
Next-Generation Storage at VMworld
VMworld 2013 was all about flash and caching, and the management of those solutions differentiated the market leaders. This year, the interesting thing to watch is how storage vendors use flash. There are some very mature all-flash arrays on the market now, as well as some excellent next-generation tiering arrays that blend the economics of near-line SAS or SATA with the speed of flash. Also interesting is the choice of most of these vendors to use iSCSI or NFS. These protocols aren’t outcasts anymore, and organizations are starting to question the need to have expensive, duplicate storage networking in the data center when they already have IP-based connectivity.
As with hyperconverged products, there are plenty of questions to be asked around deduplication, compression and compression ratios, and replication. You should also ask about scaling. Remember that storage is measured in two ways: capacity and performance. Some vendors’ arrays can scale to petabytes, others to more meager sizes. Vendors that have scale-out arrays add both capacity and performance simultaneously. Others that have more traditional, monolithic arrays have a permanent performance cap. Quality of service is also becoming a distinguishing factor in arrays, and might be of interest if you’re talking to vendors. How does an array guarantee predictable performance for a specific workload?
I’m looking forward to the conference this year, and I hope you are, too. It is always an exciting time of the year for virtualization. If you are at VMworld 2014 and would like to connect, please reach out!