There has been a lot of talk about SD-WAN these days, and a lot of questions about what to do. I have been working with a lot of clients on these questions and so I wanted to present my thoughts on the market.

Is SD-WAN ready for Prime Time?
No, not really. While it is an impressive technology, the market is still very immature. When I talk to the various vendors, there doesn’t even seem to be a consensus on what constitutes a common feature set for an SD-WAN device. Some are banking on security, others want that to be pushed to the cloud. Some tout WAN acceleration, others don’t see a need for it, the list goes on. This just tells me that the vendors have not figured out how to differentiate in this market yet and that customers haven’t really identified all of there pain points.

Another telling sign is when you look at the Gartner Magic Quadrant. There are 20 SD_WAN providers with 14 of them sitting in the Challenger or Niche Player space. What this tells me is that the market hasn’t shaken out yet. There is a lot of attrition and acquisition still to happen in this space, which makes it unpredictable and hard to evaluate.

These products are intended to replace traditional routing infrastructure, and that is not something a business should take lightly. These products are new, more complicated and likely to have some serious bugs that will impact availability and security. The networking community is notoriously cautious about these things.

Should I invest in SD-WAN?
Probably. The reality is that SD-WAN is the next generation of WAN routing. If you are due to refresh your routers, it probably doesn’t make a lot of sense to continue to invest in old technology. While we are still waiting for the market to mature, it is probably mature enough to start digging into it if you are due for a refresh. If you can afford to wait another year, the winners in this market will start to become known.

If you choose to invest in SD-WAN, you must look at it as part of a broader strategy of WAN rationalization. The greatest benefits to this technology are in the cost savings you can get out of making less expensive lower reliability circuits work with the same reliability and throughput as more expensive circuits like the MPLS networks so prevalent in most organizations. Still, you are going to be limited by what is available in each location, and at the end of the day, if the circuits to that location are challenged, there is only so much any of these products can do.

If your organization has a lot of voice or video data traversing the WAN from remote locations, these products have a lot of interesting features that help improve the quality of these services (although most have their own take on how to do this). 

Another benefit that these platforms have is that they tend to give you greater visibility into the applications that are running across your network, plus more tools to control them. There is a case to be made that, in addition to reducing your spend on circuits, you could potentially replace some of your monitoring infrastructure with some of these products. In fact, companies like Versa and Cato who focus on the security side of SD-WAN afford you the opportunity to replace your remote firewall infrastructure. While these are not very mature security products yet, and they need to prove themselves out in the field, it is a compelling case.

There are so many players in the space, who should I evaluate
Gartner got it mostly right. When you look at the Leader quadrant you have Silver Peak, Cisco and VMWare. These would also represent the technologies you see most often in the managed service provider space (Versa would also fall into that bucket). Cisco acquired Viptela, and VMWare acquired VeloCloud, so these companies are already past the market consolidation phase, and both of those companies have strong track records in the networking space. SilverPeak also has a strong track record in the networking space but their product is more home grown.

One of the concerns often raised about Cisco, is how they handle acquisitions. They will likely have a desire to see this as a technology to bolster equipment sales and will focus on integrating it into their routing and switching products, and we are already beginning to see the early stages of that. Depending on your world view, this could be a bad thing or a good thing.

The three companies in the visionaries quadrant, Versa, CloudGenix, and Cato Networks are also interesting players. Versa and Cato lean heavily on the security side, while CloudGenix has an impressive and easy to use platform that certainly warrants a look. CloudGenix is also looking to forge relationships with the MSPs. These three companies have a lot going for them, but they lack the financial resources of the bigger players. While I would not discount them just for being startups, there are challenges that are easier for a large established company to solve for. 

Should I do this myself?
You can, but it might make sense to wait it out or let someone else do the heavy lifting. Like I said earlier, if you can wait, I would give it a year to see the market stabilize a bit more. If you can’t wait, be prepared for frequent patches and updates. These are rapidly evolving products, so one should anticipate a lot of change in a short amount of time. If you are used to the slow evolution of traditional Cisco infrastructure, this may be jarring.

Be prepared to encounter poorly implemented, incomplete or missing features. This will get better over time, but as I stated previously, the standard feature sets of this product category are still in the process of solidifying. I would not be surprised to see many new features being added, while others are left unattended to die. 

All this means that there will be a fairly significant learning curve to all of this. If you don’t have the staff available to take this on and a lab environment capable of running tests prior to deployment, you are likely going to struggle with this. While SD-WAN has the promise of making the management of the WAN easier, that dream is probably still a ways off.

Most companies are opting to go with a Managed Service Provider. In fact, a lot of these vendors are primarily focused on selling to the MSP market, as that is where they see their biggest opportunity. This makes sense for a lot of organizations. Let the MSP do the heavy lifting of all the patches, upgrades, testing, and troubleshooting. At scale, they can likely do that more effectively than you can in your enterprise. It will be more costly than you can do it yourself, and some of these providers will no doubt be frustrating to work with, but one must count risk not just money in your calculations. Additionally, many of these providers withhold access to some of the features on these devices that they find problematic to manage, so if you must have the latest and greatest instantly, this might not be the best path for you.

For those who are looking to replace circuits, but are constrained by spend commits with the telcos, it might be worth looking at their managed SD-WAN services as a way to offset some of that commit.

Conclusion
The market is still maturing, this is an evolving and maturing product segment. While the nature of networking is conservative, there is enough going on with these products to warrant your attention. If you are ready to refresh your routing infrastructure, it may be a good time to invest in SD-WAN. Make sure you are prepared to take on the broader WAN rationalization project or you won’t see the greatest value. Think about leveraging an MSP to give you, and the market time to figure things out.