Catch up with the rest of our VMworld 2014 coverage here.
VMworld 2014 is in the books. It is the one event many look forward to on their yearly conference schedule; the event that provides an interesting mix of applications, servers, storage, virtualized OS (of course), networking, cloud services, cloud management, and just about everything in-between. Maybe it is just me, but I missed the major excitement this year. No huge announcements, no ground breaking solutions or industry alignments.
Here’s my take on what did and didn’t catch conference attendees’ attention during VMworld.
The tone of VMworld is set by the announcements VMware makes leading up to and during the conference. VMworld 2013 was buzzing about the Nicira acquisition, and you could feel its influence throughout the entire event. It’s now a year later, and there is still very limited deployment of that solution. In VMware’s defense, this is not simple technology. Networking is still far more complicated than it should be, even when you virtualize it into an overlay network.
Of note, NSX has made great strides to remove the complexities from its solutions. Specifically, they are abstracting the work to a point where the “server guy” can do most of the network administration. However, I think there’s still some work to be done there in terms of the physical placement of routing points and redundancy.
NSX and OpenStack
VMware announced version 6.1 of NSX at VMworld last week, and while none of the new functionality is earth-shattering by itself, there are some interesting patterns of strategy, direction, and vision that can be gleamed from its functionality.
While version 6.1 is advertised as a microsegmentation solution (able to apply firewall and other policies to small groups of virtual resources), it really signals the start of what is a long and overdue path toward policy management and application. Across all physical and virtual components of a typical network, policy definition, application, and management seem to have always been a few steps behind. We have security, load balancing, and other applications that have decent policy engines and even abstractions, but have still been left with the management of individual components or functions.
Take into account the microsegmentation capabilities of NSX 6.1, add to it the rather active partners that are heavily policy-based, such as F5, Palo Alto Networks, Rapid7, etc., then look at VMware’s active role in the development of Congress for OpenStack. You can perhaps conclude that the entity that controls policy across a software defined data center (SDDC) is likely in the driver’s seat for one the largest, unsolved networking problems.
This brings us to OpenStack. VMware intends to release an OpenStack distribution that leverages many of VMware’s management applications for compute, networking, and storage on OpenStack clouds. OpenStack is nowhere near the maturity level needed in order to be a viable option for the typical enterprise data center. But by embracing it and providing its own applications on top, VMware has taken a leadership role in driving OpenStack to enterprise maturity. OpenStack eventually would have gotten to enterprise maturity on its own and at VMware’s expense, meaning VMware’s embrace of its would-be competition was an excellent move.
EVO and SDDC Packaging
The final VMware move of note from the conference was its announcement of EVO, an integrated architecture for data center compute, storage, and network. Created with hardware partners, the first incarnation, EVO:RAIL, is essentially a four-host, 100-VM reference architecture that bundles vSphere, vSAN, and VMware support and services on top of a 2U server chassis from the likes of Dell, Fujitsu, Supermicro, and others. Offered as a single SKU, EVO is mostly a convenient packaging and ordering step. Its next incarnation, EVO:RACK, may be a VMware partnership-based response to VCE, Cisco Flexpod, IBM Flex, and a host of others, with Cisco being its main target.
There is no doubt the SDDC market is progressing quickly, but there are so many moving pieces that it is hard to package it in a way that the typical enterprise consumer can actually consume it. While we move away from large monolithic applications to more agile virtualized applications and resources, we will find that the need to architect, design, and implement this complexity may never be greater.
Combined with a storage- and management-heavy solutions exchange, VMworld 2014 had a bit of an “in-between years” feel to it, the same as those sports teams who find themselves needing to rebuild following an awesome season. This year there wasn’t a buzz of excitement. Not in the announcements, not in the keynotes and not on the solutions exchange floor. But, if we as an industry see this event as the most important enterprise data center event of the year, then we can (and should) do better for ourselves and those that take the time to really learn where the industry and solutions are going.