As 2013 draws to a close, there are a lot of people posting predictions for 2014. As for me, I think that 2014 is going to be the year that VDI really takes off…
But seriously, I’m not big into making predictions. If I’m right, there is just more pressure to do it again next year, and if I’m wrong… humiliations galore. Instead, I’d like to look at where we have been as leading indicators to what may be interesting going forward. To that end, while I am migrating labs between clouds in preparation for VMware Partner Exchange in February, I’m looking back at 2013 and some things that piqued my interest. I’ve tried to keep the list short because you’re probably like me and have a short attention span. There are so many to choose from, but three is about all I can manage.
For me, 2013 was a pretty big year: I joined VMware in March, attended my 9th and 10th VMworlds (including my first one in Europe), drafted my first white paper, and learned quite a bit about NSX, VSAN and OpenStack along the way. That’s not to say that I’m an expert in any of those areas yet, but I think each of them is going to be important in the not-too-distant future. You may think that I’m just drinking the company Kool-Aid, but the engineer in me who likes to look at the components of a solution thinks these technologies are actually very interesting.
NSX
We officially announced NSX at VMworld San Francisco in August of 2013 and its GA in October in Barcelona. Those of us on the Hands-on Labs team witnessed the excitement around these announcements at both conferences: the NSX for vSphere lab was taken more than double the number of times as the second-most popular lab (our usual winner, the vSphere Performance lab.) I think this is the first time that we have had such a dramatic difference between our top labs.
Over the past few months, the interest in NSX has not waned. Within the public Hands-on Labs, we have delivered nearly 2,500 NSX labs since October. People want to understand this new paradigm and how they can leverage it to make their lives easier. Don’t get me wrong, the transition is not something that can be done overnight or without planning, but I think the potential value of this abstraction for the right customers is incredible. If you would like to experience NSX for yourself, we have two Hands-on Labs for that: HOL-SDC-1303 (NSX for vSphere) and HOL-SDC-1319 (NSX for Multi-Hypervisor). In HOL-SDC-1319, you create each of the components and connect them by hand using the web interface. In a later module, you have the option of performing all of that same work by executing one simple script. There is a lot of untapped power in that kind of automation.
I have not focused in the networking space for many years and I am looking forward to reacquainting myself. I know it has been said by many others before me, but NSX seems to be positioned to do for networking what ESX did for servers in the early 2000’s. Watch this space.
VSAN
I’ve worked a lot with enterprise storage, so I think this is a no-brainer. That I’m talking about VSAN may surprise you since it involves decentralizing the storage. However, I’m also a big fan of choosing the right tool for the job, and I think there are jobs at which VSAN will excel. Even though VSAN is only in beta, there is a great deal of excitement about this product. I have been in standing-room-only VSAN sessions at VMUG meetings, and nearly 10,000 people have expressed interest by signing up for the VSAN beta program — you can go to http://vsanbeta.com and join the party.
Even if you have a “proper” storage area network with a fancy array, I humbly suggest you take a look at what VSAN can do and, more importantly, how it ties storage and virtual machines together. Beyond its underlying host-centric storage configuration, which may give traditional SAN people that funny feeling in the pits of their stomachs, the per-VM storage policy based management (SPBM) is absolutely worth understanding. SPBM has applications beyond VSAN. Say you are interested in VSAN but lack the gear to get it installed. You can head over and take the VSAN lab in the Hands-on Labs. While you will not experience the performance of running a real VSAN, you can have a look at the interface, enjoy its configuration simplicity, and explore those storage policies I mentioned earlier.
OpenStack
OpenStack is another thing that caught my eye this year. There has been all kinds of buzz around this one, and I admit that I initially looked at it because I wanted to know what all of the fuss was about. If you are looking for a way to deploy certain types of applications or workloads, this framework is absolutely worth a look. However, I suggest a nontrivial amount of Linux knowledge as the prerequisite to any production deployment.
We have been doing a lot of work with OpenStack at VMware — in fact, VMware is one of the top 10 contributors to OpenStack, and no small effort has been expended to ensure that OpenStack integrates not only with VMware vSphere but also NSX and VSAN. Some people believe that OpenStack and vSphere are mutually exclusive. That is entirely untrue. You can learn more about all of this on our VMware OpenStack community page.
I noticed a lot of confusion in the industry regarding OpenStack: what it is and how it can or should be used. As someone who comes from a traditional datacenter background (you know, 1:1:1 application:server:OS, or application:VM:OS), I think I am of the same mind as many of you. I have done my fair share of research and I see OpenStack as an interesting technology that has its place in the toolkit.
Unfortunately, I believe there are some in the industry trying to position OpenStack as the “one ring to rule them all.” It isn’t, at least not out of the box or without accepting a significant amount of risk. If you have heard the pets/cattle or kittens/chickens analogies before, you have an idea of what I mean. At this point, OpenStack is most definitely suited for “cattle” and “chickens,” and it does a great job with those. Users with the right workloads seem to love their OpenStack. However, I have not seen it used to provision Microsoft SQL or Exchange servers, Oracle application or database environments, or most traditional (i.e. not cloud-aware) enterprise applications.
When I first read about OpenStack, I thought, “Oh, great, is this the the ‘single pane of glass’ panacea all over again: manage multiple different stacks of gear using just one API!” After closer inspection, I think there is more to it than that and it is worth watching to see how deep the rabbit hole goes. As long as it somehow manages to keep things simple and doesn’t contract “creeping feature-itis,” I see OpenStack and its offspring going a long way. Interested in some Hands-on? Yeah, we have a lab for that, too: HOL-SDC-1320 (OpenStack and vSphere) NEW Openstack HOL available -> https://my.vmware.com/web/vmware/evalcenter?p=openstack-hol
In my experience, it is good to know what the available tools can do, even if you do not yet know all of the details about how they work. Sure, I could have easily mentioned the cool stuff that the PernixData guys are doing, EMC’s ViPR, our own vCloud Automation Center, or a multitude of other intriguing solutions out there. Those other things are all very interesting to me, but there just is not enough time in the year to research everything and get my day job done. So, I chose two of our products and an open source project because I invested enough of my time in 2013 to understand what they mean and how I see them being applied in the future.
Cheers and a happy 2014 to everyone. Take a look, take a lab!
