The hybrid WAN is an idea that’s been around for decades now. Prior to being an analyst, I worked in corporate IT and I was building Internet VPN links to connect home office workers, small branch offices, or back up connections. Back in the “old days”, a hybrid WAN was definitely the exception and far from the norm. One of the companies I worked for had an old school CIO who couldn’t accept the fact that an Internet based VPN could actually out perform a private connection, because it wasn’t “guaranteed”.
Today, though, the concept of the hybrid WAN is gaining acceptance. However, I still feel there are a number of truths that need to be accepted before we see hybrid WANs become the norm instead of the exception. Below are the top 5 “leaps of faith” that need to be overcome:
- Guaranteed performance isn’t always better. Conventional wisdom dictates that the guaranteed performance of a high price MPLS connection must be better than a non-guaranteed connection. Makes sense, right? Well, not exactly. The performance of a connection can be guaranteed because the network operator offers an SLA that it knows it can hit. Internet pipes can give performance that’s just as good or better than a higher price connection, it just can’t be guaranteed.
- Internet connections can be a primary link. If businesses are using an Internet connection, it’s most likely a backup link, even though the speed may be faster. I’ve talked to some organizations that use the Internet connection as the primary link because users notice a performance boost when the primary fails. If nothing else, businesses should route all non mission critical traffic over the Internet and use the MPLS connection for voice and video traffic.
- Quality of service can be used for more than just private connections. QoS is one method of guaranteeing the delivery of traffic. It’s primarily done via a two sided solution where a “queue” is created between two points. While you can’t exactly create a queue out to the cloud provider, the technology does exist to do some traffic shaping over the Internet. Also, with the rise of NFV, it’s possible to deploy a virtual appliance and implement QoS directly at the cloud provider.
- Multiple paths can both be active. For some reason the network industry lives with this fallacy that any secondary connection must be passive. The “active-passive” model when building a WAN is what the overwhelming majority of companies do today. This is akin to building two roads between two cities and having all the traffic use only one of them and only opening up the second when the “active” road fails. There are a number of great multi-path products available today to enable multiple paths to be active at the same time without the risk of routing loops.
- Building a hybrid WAN isn’t complicated. Years ago, the implementation of a hybrid WAN was very difficult. One would need multiple appliances for routing, security, backup links, etc. All in all, the rewards of a hybrid WAN outweighed the benefits largely because of deployment complexity. Not so today. WAN Optimization, routers, security and other network devices have evolved tremendously giving network managers a cornucopia of options to build a hybrid WAN without the risk and complexity that existed a decade ago.
The hybrid WAN is here to stay and early adopters will enjoy a better performing WAN at a lower cost. For those of you luddites (skeptics) out there, in more circumstances there will be little to no performance drop off at a significantly lower costs. All you need to do is take a leap of faith.