Prediction is very difficult, especially if it's about the future. - Niels Bohr
2015 is well under way, and while I have always eschewed writing anything that even smells predictive, I’ve made an exception this year. Perhaps it’s because the networking industry has been changing so rapidly that I can’t help but be wrong, so what do I have to lose? Or perhaps it’s just for an excuse to work Niels Bohr into an article, making me feel better about myself as a person.
Layer 1 Switching
Layer-1 (or Crosspoint) switching is less a new technology and more a repurposing of very old technology. Think the old telephone networks’ switchboards, where electrical connections were manipulated by humans based on the needs of the network. The cabling largely stayed the same on the backend. Layer-1 switches allow for the permanent connection of all layer-1 (wiring) connections across the datacenter, while certain manipulators like OpenFlow and Python can “move” those physical connections logically to allow for a re-cabling of the network without actually having to re-cable anything.
While not deployed by many networks yet, look for this technology to emerge as one of the must-haves in the next year. Even if still not widely deployed by the end of the year, a lot of enterprises have started to take note of the myriad possibilities made real by the flexibility of cabling datacenters using Layer-1 switching and orchestration products by vendors such as Qualisystems and Metamako.
Lab virtualization is a huge and growing application for this technology, as are flexible and large-scale datacenters. This isn’t a cheap or easy technology to deploy correctly, so I don’t expect to see this in any but the largest of enterprise networks and labs. Still, as knowledge around this builds, prices and the overall barriers to entry will drop. By the end of the year, look for more companies to be doing proof-of-concept testing with some of this gear.
ACI vs. NSX
This will continue to be the dominant developing story in the SDN world this year, at least in the enterprise space. Both vendors have invested heavily in their respective technologies, and neither has shown any sign of slowing down in 2015. VMware has come out of the gate close to a year ahead of Cisco, and while they have had some success, a large portion of the enterprise space is waiting for hands-on with ACI—something just now happening—before committing to any solution.
Geneve, VXLAN, STT, NVGRE
The Geneve spec, currently in a draft proposal status in the IETF, is interesting in that it combines elements from VXLAN, STT, and NVGRE as well as being backed by Microsoft, VMware, Redhat, Intel, Broadcom, Brocade, Arista, and Cumulus. Notably absent are Cisco, Juniper, and Avaya, but with such a group of heavyweights backing the draft already, adoption may be a foregone conclusion. We’ll continue to see development in overlay technologies like this, much like the VLAN and MPLS technologies which came before them. It remains to be seen where we’ll finish the year, but this emerging technology has the smell of inevitability about it. It feels like it is going to put pressure on vendors like Cisco to eventually support it, especially as the latter makes moves towards a more cozy relationship with Microsoft.
The OpenDaylight project will continue to be an important driver in the mindshare of the open source and SDN worlds, while at the same time not necessarily putting vanilla product into the hands of customers. By that, I mean that many companies are releasing technology based on the ODL project, but with proprietary twists to suit the particulars of that vendor. The ODL is much like Linux itself—it drives a lot of other products that use it, but nobody does it the same way. Purists bemoan this, but purists don’t have mindshare in the office of the CFO, where pragmatism tends to trump ideology.
With all of the emerging standards, methodologies, and technological shifts in the industry, it should surprise no one that existing allegiances will be tested, strained, and outright broken. Cisco and VMware are moving from “frenemies” to full-blown competitors; Microsoft and Cisco are moving closer together, VCE bought back a lot of its stock, and VCE and EMC continue to experience channel conflict behind the scenes. Rumor has it that Cisco and NetApp are working toward releasing a FlexPod based on Microsoft. OpenStack is challenging VMware on the private cloud front and AWS and Azure on the public cloud front. Cisco’s foray into storage—despite claiming otherwise—looks to be an attempt to lock-up more of the stack, while more and more storage vendors are making serious pushes into compute. HDS has been gaining mindshare with their compute platform, despite misgivings from a lot of corners of the industry.
OpenStack continues to evolve, and as it does a greater number of companies are casting weary eyes at their existing vendor bills and wondering if this isn’t a better route. The number of consulting engagements I’ve had where OpenStack proof-of-concepts are being deployed is huge, and I expect this to continue to grow significantly. As it does, look for amended pricing models to come from the more established players in this market.
Public cloud vendors are not cheap, and the more the C-level executives get eyes on the bills from these services, expect the use cases for full-blown private clouds using OpenStack to gain momentum. Running an OpenStack private cloud, especially with automation technologies for both backend provisioning of storage and virtual machines coupled with catalog services for applications, is going to become even more popular than it is today. As with other cutting-edge technologies, as more and more people become familiar with the ins and outs of deployment and management, the knowledge base will grow linearly, with adoption rates following closely behind.
Public Cloud: Azure, AWS
On that vein, public clouds are going to continue to be a popular option for capacity bursting, but I see a contraction coming. The prices as they are today are not tenable for anything but development work, short-term testing, and the very largest enterprises with no good visibility into IT OpEx. Prices are going to have to come down to maintain market share, but with OpenStack nibbling around the edges, I still don’t see long-term viability for this model in the capacities we’re seeing today.
IT automation is a hot topic these days, and for good reason. The driver behind many of today’s hottest technologies, from virtualization and private cloud, to overlay networks and software-defined abstraction of all types, is primarily flexibility and speed to deployment. As automation and abstraction have changed compute and storage significantly in recent years, the network has lagged behind in several key areas. As a result, expect to see the serious disruption in this space continue to accelerate and bring uncomfortable changes to the status quo. While I see the possibility for industry heavyweights to be put back on their heels a bit, I don’t think the trend will last, as they either come back hard through R&D monies or through outright acquisitions of key technologies.
Storage: All-Flash Arrays
The all-flash array market is heating up something fierce, and many of the established vendors in the storage market aren’t necessarily the ones on top today. Upstarts like Pure Storage are disrupting the big-boys. I see that trend continuing in 2015. Meanwhile, don’t count out the established enterprise players like EMC, NetApp, and HDS as they make aggressive moves to take back what they haven’t had to fight for in a while. I don’t see the overall tier structure between storage vendors like HP and Dell changing this year as they stay in the mid-market space and don’t seriously disrupt the enterprise players much at all.
I could easily write a few more pages of predictions, if for no other reason than that there’s a heck of a lot going on in the IT world these days. So let’s just leave off where we are, and if feel I’ve missed something major, or you’d just like to point out places where you disagree, feel free to leave your comments below. I look forward to reading them and to seeing how much of this pans out over the upcoming year.
Photo courtesy of Shutterstock.