The Cranky Admin
Virtualization Predictions for 2017
Two big trends: Microsoft becomes the hybrid cloud leader, container wars turn bloody.
As years go, 2016 was dreadful: politically, economically, and even in terms of excitement in the storage and virtualization segments of the IT industry. The past year saw the start of a storage crisis that will bite us in 2017; yet another iteration of various products; and some mergers and acquisitions. Even though that sounds like a lot, the reality is that 2016 was just boring!
This year promises to be a bit more exciting. For one thing, the feud between VMware and Nutanix has entered a new phase. VMware is now overtly running scared, while Nutanix is out there buying up the pieces it needs to deliver a proper turnkey cloud-in-a-box solution; something VMware has tried a few times to get right, but consistently failed on.
The past year saw VMware largely give up on public cloud, ceding the space to Amazon and accepting a background role as a purveyor of "hybrid" cloud software that relies on larger players to provide the public cloud side of things. This year is where we get to find out if this strategy means VMware sees renewed relevance, or becomes merely a hapless funnel through which workloads pour from on-premises datacenters through to the public cloud.
Azure Stack solutions should start shipping in volume this year. Windows-based containerization and the next generation of Hyper-V hypervisor will also see large production deployments this year, having both come out late in 2016.
The Microsoft Gorilla
If the release candidates are anything to go by, 2017 will see Microsoft go from virtualization also-ran to the 800-pound hybrid cloud gorilla with no viable challengers. Microsoft has nailed the ease of use thing, has its own respectable public cloud, and has answers for those needing containers, hypervisors, what-have-you-as-a-service, cross-site and cross-cloud connectivity. All of it can be managed with PowerShell and most of it can be managed through Azure Stack. Expect 2017 to see announcements about finalizing the integration of the last management pieces into a single Azure Stack interface.
This will also be the year of the turnkey cloud. Several smaller players, ranging from Yottabyte (disclosure: they are a client) to Zerostack to Nodeweaver, already exist. There are also easy-to-use software solutions such as Platform 9 and Stratoscale, and virtual appliance solutions such as vOneCloud.
What's important here is that the existing solutions are reaching levels of maturity that make them ready for production. At the same time, hyper-converged infrastructure (HCI) players -- most notably, but not limited to, Nutanix -- have clued in to the bit where those that don't provide a turnkey cloud solution may not survive to see the end of the year.
Similarly, as the battle moves into the multitenant and self-service amendment layer, the hypervisor largely ceases to matter. KVM is already popular amongst HCI vendors and is only going to continue to attract customers. The current year will close with Google not being the only one using KVM at scale.
Bloody Container Battles
Containers are starting to not suck. We're now at the point where, if you know what you're doing, you can use them for either static or ephemeral production workloads. Just try to avoid the holy war about how they "should" be used. Now the bloodshed can really begin, as the technology ecosystem starts to thrash out who makes the best toys to make containers easy enough for mortals to use. Companies with truly novel solutions, like Kaleao, will be interesting to watch here.
The public cloud is entering a new phase of competition. The channel is terrified, and there are various interesting experiments running to see just how channel companies should reinvent themselves in the face of this change. There are a lot of companies out there looking for reasons we should pay them money, and the deals to be had are almost heartbreaking.
Ignoring Google and IBM
Amazon and Microsoft have settled into a cold war, with Microsoft fixated on acquiring market share by cannibalizing its channel and Amazon taking the "spam more features" route. Both are so consumed with one another that they've functionally ignored Google and IBM.
Google still hasn't decided what it wants to be when it grows up, and thus remains the public cloud of academics and people who want to do things that can only be described by and to those with math degrees. IBM, on the other hand, has been thoroughly boxed about the ears by both its customers and the channel, and is altering its approach to the public cloud in response.
I personally think IBM will close out 2017 as a serious threat to both Microsoft and Amazon, leaving both of these companies wondering how it happened. In the meantime, there are some great deals to be had for larger public cloud customers in playing the "big four" vendors off one another.
The Rise of Adaptive Security
One thing I expect to see make a sharp rise to prominence that hasn't been widely discussed to this point is adaptive security. Think virtual appliances that are hybrid cloud-based, Big Data backed, make use of machine learning and plug into our hypervisors, microvisors and network controllers.
These solutions will analyze traffic and I/O patterns in real time, matching against signatures of known baddies, employing blacklists and using heuristics to look for things that stray too far from normal. The goal will be to not only provide alerting, but automated quarantine, imaging, data isolation, in-depth tracing and forensic capture. Automated incident response systems like this already exist, but they've been pretty niche until now. Expect 2017 to be the year they emerge from infancy and start showing up in our datacenters.
WAN Overlay Networks
The last item we need to look at is WAN overlays. A new class of WAN accelerator is emerging where you put an appliance (physical or virtual) in your datacenter and pump all your traffic through that appliance. The appliance then determines whether to route your traffic through your MPLS connection to another site, out onto the raw Internet or through a tunnel to a WAN overlay network.
The WAN overlay networks consist of both virtual routers sprinkled throughout the Internet, and a whole lot of data capture and analysis. In short: they scan the Internet to see what traffic patterns look like, then route your traffic through their network of virtual routers to avoid congestion.
Companies to watch in this space include Dispersive Technologies and Teridion, though there are many others. Each has their own unique approach, and all promise to make our access to public cloud services suck less.
Clash of the Titans
Overall, there's a lot of excitement to be had in tech in 2017. Some new segments emerging provide some bright cheery spots, while the rest of the entertainment is to be had watching various titans collide.
About the Author
Trevor Pott is a full-time nerd from Edmonton, Alberta, Canada. He splits his time between systems administration, technology writing, and consulting. As a consultant he helps Silicon Valley startups better understand systems administrators and how to sell to them.