SD-WAN Service: Keep Your Eye On The Carrier Bottom Line

All too often in an SD-WAN evaluation, a client at some point will ask about purchasing an SD-WAN service from one of the carriers. Now, I’ve got nothing against carrier-managed services. After all, MPLS has been the mainstay of enterprises for years, and before starting SASE Experts I ran MPLS-Experts.

But at the same time, for years I’ve heard customers grumble about MPLS providers. Taking 90 days to receive a new MPLS circuit became harder to explain to the business when executives receive xDSL at their homes within a few days.

The bandwidth spend on MPLS also becomes harder to justify when internet bandwidth is half, even a quarter, of the cost.  Granted, the Internet of 10 years ago wasn’t what it is today. There was far more loss and latency on even local routes. All of which helped make the case for staying with MPLS. But as the internet’s quality improved, companies are questioning, “Why MPLS?”

All of which makes shifting to carrier-managed SD-WAN particularly ironic. The technology practically invented to address limitations of carrier-managed services is now going to be provided by these very same carriers? (Take a look at our explainer so you better understand how SD-WAN meets MPLS limitations.)

I think not. SD-WAN’s ultimate purpose is to give us the freedom to use any available data service from any service provider. But how much freedom can you have when the carrier has a vested interest in keeping customer locked into their network?

To put that another way, businesses will always pursue their financial interests.  Carriers are no different. MPLS services represent a cash cow. They simply can’t be trusted to slaughter those cows for the sake of some new technology.

And it’s not just me saying this. Carriers themselves are acknowledging the issues. Customers aren’t replacing MPLS with the internet as much as they are augmenting existing internet connections, noted Verizon’s Senior Vice President of Business Products Shawn Hickey in an interview with SDXCentral “What we see is a lot of MPLS plus LTE backup.”

It’s true that part of the reason customers can’t leave MPLS with SD-WAN is because of the internet’s quality. While local internet routes have become far more predictable, global internet connections, particularly from North America to AsiaPac,  are often very erratic. Depending on the application, it’s criticality to the business, and user expectations around quality of the experience, the internet is often unsuitable exclusively for global enterprise transport. All of which provides a rationale for carriers keeping their customers on MPLS. (Even that’s not exactly in the customer’s best interest as you now have global backbones, such as provided by Aryaka Networks and Cato Networks at a fraction of the cost of MPLS with very similar performance.)

We’ve discussed the problems of the global internet before as have others. What’s less discussed, though, are the business drivers carriers face when selling SD-WAN services. As the cash cow, there’s a big disincentive for carriers to eliminate or reduce MPLS capacity. Even though they’re switching from MPLS to SD-WAN and not losing the customer, the profits from an SD-WAN (with internet) service is a fraction of MPLS.

“….a SD-WAN site service can be 5 to 8 times cheaper than the MPLS counterpart, depending on the particular site circuit status, access configuration, and class of services (CoS). This means when a client migrates their WAN service from MPLS to SD-WAN, the MPLS service provider will either lose the account to other vendors, or erode the current revenue to a much lower level despite of the rescued client’s loyalty,” David Wang writes in Telecom Ramblings. Wang is a telecom/IT business development principal and senior consultant with ITCom Global.

All of which makes me skeptical around carrier-managed SD-WAN.  Even if the acquisition costs sound reasonable, I would suggest you probe very carefully to find the hidden “gotchas.” Are they delivering the same service levels? Where are the lock-ins? What are the minimum number of carrier circuits they will allow? What is their support staffing level?  Ask a lot of tough questions (or contact us for a free consult, if you don’t know how). Because any partner working against their bottom line is probably not going to be very good for your bottom line.  

Share this post