Feb 16, 2017
The Gartner Magic Quadrant is certainly the most iconic graphic ever created by an industry analyst firm. Second on that list is the Gartner Hype Cycle. Personally, I find the naming scheme of the elements of the Hype Cycle to be a bit trite, but each phase is fairly accurate.
Technology innovation cycles start with a group of innovative, smart people that come up with a solution to a burning problem. If it’s a real problem, it attracts a number of established vendors and start-ups that quickly hop on the bandwagon to create a real market. If not, the market quickly vanishes. If the market is real, vendors, VCs, and the media quickly pile on and the hype will turn to mania as the market reaches the “peak of inflated expectations”. Typically, what happens in these instances is that the ROI and the “solve all the world’s problems”-promises don’t materialize and the technology goes into the dreaded “trough of disillusionment”. Once the best practices are established and the technology matures, the technology will go mainstream in the “plateau of productivity”.
Many disruptive technologies have gone through this cycle and new markets have been established. As an example, VoIP initially was supposed to create a whole world of apps that people run on their desk phone – that never happened. The complexity was overwhelming and many organizations backed off. Eventually the hard stuff was figured out and now VoIP is commonplace. A more recent example is data center focused Software-Defined Networking (SDN). Early in the cycle, the technology promised to commoditize all hardware and make the data center network an agile, flexible fabric. SDNs can certainly do that but the complexity in getting there is so high that most companies have kept it in their test labs. In fact, I’ve heard some industry people joke that “SDN” stands for “Still Does Nothing”. It’s currently in the trough and I expect that in the coming 12-18 months it will emerge and hit the plateau.
Software-defined WANs (SD-WANs) have been around a few years now and are reaching the peak of inflated expectations. However, I predict that SD-WANs will skip the typical stop at the trough and go directly to the plateau of productivity when the technology goes mainstream. There is precedence for this. For example, WAN optimization started off like a rocket and propelled directly to mainstream adoption. The market has certainly matured and slowed down, but that’s because it’s been so widely deployed.
There are two reasons I feel SD-WAN will buck the trend. First, WAN transformation is long overdue. Prior to my career as an analyst I was in corporate IT, and one of my many roles was running a global WAN. Even back in the 90s we were discussing different WAN architectures to replace the traditional, static hub and spoke. The challenge was that there wasn’t a viable option to what we were doing. The underlying transport evolved from frame-relay to ATM to MPLS, but the underlying architecture remained static.
Another reason is that the legacy WAN is increasingly holding organizations back. Digital transformation is hot right now and it’s changing everything in IT. In the digital era, speed means everything and forces companies to embrace trends like DevOps, the cloud, and IoT. But how can companies move with speed if network operations lack agility? Network speeds have steadily increased, but that’s only part of having a fast network. A fast network must also be operationally fast. That means being agile and having the ability to make changes and deploy new services almost instantly. An SD-WAN enables that.
Another compelling reason that SD-WANs will not go into the trough is that it’s significantly cheaper than what the company was doing. I’ve talked to some multi-nationals that cut their WAN spend by 70% or more. Not all companies will realize such significant savings with bandwidth, but most should see their costs cut by at least 30%.
When I think about new technology cycles, they should achieve the following:
I mentioned WAN optimization before and it serves as a tangible example of the above. Technologies that don’t achieve these goals will certainly disappoint and fall short of living up to the hype. VoIP was far more complicated than anyone initially realized. Data center SDNs are often operationally much more expensive than running a traditional network. In fact, it’s often the case that most new technologies are far more complicated than the original problem they were trying to solve. SD-WANs are a rare example that achieves all three goals.
If you’re an organization waiting for SD-WANs to mature and go mainstream before you deploy, you might be too late because this is one technology that will likely skip the traditional phases in the hype cycle.