How SD-WAN is Causing the Growth of Multi-Vendor Networking

Over the past five or six years we’ve heard plenty of discussion around the slow and steady demise of monolith single-vendor networking and a shift to multi-vendor environments. Due to the rise of disaggregation, whitebox networking, and to an extent even vendor agnostic network automation, we should all be running multivendor networks by now.

But to be honest, I haven’t seen that happen very much. Even the very largest networks I’ve worked on and customers I’ve worked with haven’t adopted a very multi-vendor infrastructure.

There are always exceptions, of course, but by and large I’ve seen organizations stay pretty consistent with deploying and operating devices from a single or at least only a very small number of network vendors.

Very recently, though, as in only the last year or two, I’ve seen things start to change. I’ve seen a willingness and sometimes eagerness to look at new vendors, but it’s been within a specific context.

It seems SD-WAN has given organizations small and large an opportunity to expand their horizons and implement solutions from more than one vendor. Instead of insisting on only one brand for the entire network – the campus, data center, WAN, and even security – IT managers seem very willing to entertain peeling off the WAN and going with a completely new vendor.

For example, a WAN router refresh might include most if not all of a company’s routers especially if they’re co-termed for support or if they all go end-of-life around the same time. It’s a chance for IT managers to rethink how to manage a WAN and see what’s out there.

Instead of simply upgrading the old box with a new box – a like-for-like rip and replace – I’m seeing organizations take the opportunity to explore new options. A wide area network refresh is the perfect opportunity for a new networking vendor to slide right in.

A campus refresh might seem easier since for the most part a campus is mostly access switches. However, think about it in terms of budget. A network of several thousand access switches to go from vendor A to vendor B in one big refresh would be obscenely expensive and take years to implement. And I just don’t work with many IT directors and network admins that are OK with a mixed vendor access layer.

A WAN edge is far fewer devices overall, and since it’s a different logical part of the network, I’ve seen network admins much more amenable to a different vendor for the WAN than they use for their campus and data center.

Today, I have Cisco-only customers considering Velocloud for their WAN while keeping Cisco as the primary vendor everywhere else. I also have anything-but-Cisco customers interested in deploying Cisco ISRs with SD-WAN code because of the built-in branch security and granular control available in vManage.

And this openness to a new vendor is compounded by the fact that many organizations now service chain their new SD-WAN solution with yet another vendor’s cloud-delivered network security solution.

No, this isn’t the same as pure multi-vendor in which the brand is completely irrelevant anywhere in the network, but it’s certainly a step away from full-stack single-vendor networking enterprise IT has been enamored with for years.

It’s been interesting to see the adoption of SD-WAN open the door for new networking vendors to get a foothold in what were always single-vendor environments. It’s not necessarily that whitebox and disaggregation panacea that the industry has been touting for a while, but in my experience it’s certainly a very real shift to a <somewhat> more multi-vendor environment.

Thanks,

Phil