The push for software-defined networks is unmistakable. The buzz in the industry has reached a fever pitch, and some vendors are already well en route to delivering the next generation of networking products. And in between the vendor rallying cries, if you listen closely to the background din of battle, subtle themes are emerging. These themes, while falling short of actually driving the future of networking, will play a key role in how SDN emerges. Among the most powerful of those networking subtexts is vendor lock-in.
We all know the ill effects of lock-in. Lack of leverage when negotiating price strikes fear in the heart of every procurement office across the globe. Customers whose architectural decisions have left them without an alternate plan forward can expect to see their sales discounts shrink. The negative side effects of being locked in do not end with price, either. When the barriers to change become large, they effectively commit IT departments to a particular technology track. With only a single technology partner, this effectively limits the source of innovation to a single R&D team, making the company dependent on the priorities and developer ingenuity residing under a single banner.
So how should customers view vendor lock-in? At its most basic level, vendor lock-in occurs when organizations cannot replace one vendor's gear with another's. Accordingly, the definition of vendor lock-in hinges on the concept of interchangeability. Vendor lock-in exists wherever there are not two or more solutions that are functionally equivalent. Keeping this definition in mind, here are a few practical tips for identifying and mitigating unwanted lock-in.
First, though, bear in mind that lock-in is not inherently evil. It just is. Sufficient value can justify a drop in flexibility. If you are never going to switch to a different solution, the switching don't really matter. Ultimately, lock-in is just another tradeoff, not unlike price, to be considered before purchasing anything.
With that being said, don't get locked in because of price advantages. Some people make short-term purchasing decisions because of attractive acquisition costs. These are not a reason to get locked in. Account managers know that they can create a barrier to change, so the first purchase will always come with the steepest discounts. You'll be amazed at how fast those discounts vanish once you're hooked into a particular solution or architecture. Additionally, at least in the networking space, an increasingly competitive landscape will drive prices lower over time anyway. The price advantages of today are not necessarily sustainable, so plan accordingly.
If you bet on technology, at least look forward. Some technologies provide business value significant enough to justify future lock-in. If a solution is truly that right for your business, make the decision and never look back. But before you place your bet, ask: Is the technology new? If it is, and it takes off, other vendors are likely to pick it up, which would lessen the lock-in over time. If it is old, then it is exceedingly unlikely that vendors will suddenly change tack and support it.
Lock-in is not just about features and protocols within the networking domain, but about the larger IT ecosystem, too. Networking solutions have to interoperate with a lot more than just other networking gear: provisioning systems, capacity planning tools, monitoring and troubleshooting applications, and even peripheral IT systems like help desk ticketing systems. If you do not consider the impact to all of these surrounding systems, you might get bitten. In general, the larger the number of things that plugs into a single point, the greater the lock-in potential.
Ease of integration matters. Whatever solution you create will require integration with surrounding systems. Easy and repeatable integration lessens the threat of lock-in. If integration is difficult or handled as customized one-offs, lock-in will be greater. As a rule of thumb, think carefully about any integration requiring professional services. In these situations, the integration is difficult, and once the services engagement ends, you lose access to the expertise. Customers should prefer solutions designed explicitly with integration in mind. A simple messaging check to see if integration is a top-tier priority should provide some indication.
Where you draw the boundaries matters. Buying behavior in networking, despite all the talk of solutions, remains largely on the box side. When evaluating the extent to which one solution can replace another solution, consider whether you are looking for box-for-box interchangeability or you whether you view the entire solution as a single entity. Box-by-box replacement, while often tougher to find, offers less expensive replacement routes. Swapping out an entire solution usually costs more, but will make interchangeability focus more on things like layer 2 and 3 interconnect rather than internal fabric mechanics.
And don't underestimate the people factor. Perhaps the biggest barrier to change is not cost, but rather human psychology. Once you have trained a workforce to interact with a certain element, it becomes difficult to change the interface. This gets to the very core of how people operate. As you consider solutions, determine how integral they are to people's day-to-day workflow. Solutions that are built into the process and require a lot of intimate connection between user and product will prove harder to displace when the time comes. This is not to say that user interface is unimportant and that people should choose poor tooling, but that you should carefully consider where the point of control is. If a single system acts as a point of control for many systems behind it, that front system will become increasingly difficult to change.
Ultimately, it is both impossible and unhealthy to avoid all vendor lock-in. Progress cannot be made if consensus is a prerequisite. Data center architects will need to determine for themselves whether the value is worth the lock-in that comes with it. But by being strategic about how much lock-in is tolerable and being complete in your assessment of where lock-in resides, you will be in the best position to make an informed decision that balances value and future flexibility.
Michael Bushong is currently the vice president of marketing at Plexxi, where he focuses on using silicon photonics to deliver SDN-based data center options. Prior to joining Plexxi, Mike spent 12 years at Juniper Networks, where he drove Juniper's SDN strategy, including product plans around OpenFlow, path computation element, application-layer traffic optimization and BGP traffic engineering. Prior to Juniper, Mike worked in the ASIC design tools industry.