SD-WAN

I think SD-WAN is a really cool technology that builds on the strengths of the public network which affords rapid innovation. Of course we’ve always had this possibility using IPSec tunnels instead of a private MPLS network, but now we have a centrally managed, fully automated method of offering much higher levels of service than before. Application-level routing, prioritization, security and all important analytics are basic components of almost all SD-WAN platforms.

What I personally think is so great about SD-WAN is how each offering has the ability to do something special and unique to them. The lack of standards (yes I know MEF 70 exists) I think is actually a good thing in that there are no rules to follow. Each platform can do their own type of analysis, their own routing algorithms, their own prioritization, and methods of security.

Another really cool thing is that the software can run on a variety of hardware and does not need specialized proprietary boxes. The software can run on a virtual machine in the data center or remote site, or even on a virtual appliance in the public cloud. In a very short amount of time, sites can be connected securely to private data centers or to the private side of the public cloud. Gone could be the days of 30-60 day lead times for circuits and weeks to provision and turn up services. Once broadband or Dedicated Internet Access is secured, a site can be connected in a matter of minutes.

I’ve read many papers and posts stating 2020 is the year of the SD-WAN. I think that’s largely true. Of course, with a suddenly near universally remote work force, where does this leave the need for office connectivity? Have we passed SD-WAN and should now focus on software defined perimeters?