If you're responsible for a serious enterprise WAN, do you need MPLS?
Until recently, the answer to this question for almost all larger enterprises, many mid-sized and some smaller ones as well, was a resounding yes.
Thanks to the Next-generation Enterprise WAN (NEW) architecture, the answer going forward is: probably not.
Now, might there be CYA reasons why many more conservative managers will choose to keep, and even expand, their expensive MPLS indefinitely? No doubt. Having written last year that I'm a big Geoffrey Moore fan, Moore reinforces that in the Technology Adoption Life Cycle on the back half of the curve there are the late majority and the laggards. Because there are a lot of latecomers in adopting new technology, that means roughly half of the market is still likely to be using MPLS as their predominant WAN technology several years from now.
As I wrote then, my targets for this column are the early adopters and early pragmatists in deploying newer IT technologies. I expect almost all of them, and a large fraction of the early majority, however, to deemphasize their MPLS spending – with some actually decommissioning MPLS altogether – given the myriad benefits that the NEW architecture delivers, combined with the lure of public and hybrid cloud computing.
And while there are some early adopters who have already given up MPLS altogether (for example see here, here and here), most enterprise WAN managers have held back on moving away from MPLS until recently. And they have had some good reasons, together with some that, frankly, weren't so good in the first place.
Here are just some of the reasons people will cite for why they need MPLS:
- "I need a full mesh between my locations because that's how our traffic patterns are"
- "I need a full mesh for VoIP"
- "There is no QoS on the Internet, so I need MPLS"
- "Packet loss and jitter on the Internet are too high, so I need MPLS"
- "I can't use asymmetric DSL; it's not workable for our network"
- "The MTBF and MTTR of broadband connections are too high, so even if they are lower cost and higher bandwidth, I can't afford to use them"
- "MPLS provides me with SLAs; Internet connections don't"
- "To get reliable, predictable performance, I need MPLS"
- "I need MPLS for VoIP"
- "I need a single place to point the finger when something goes wrong with the WAN"
I consider the above, and other similar "reasons," as "myths" about why enterprises need MPLS. I write "myths" because some of these were never really true, and others are now addressed by the combination of technologies and capabilities that the NEW architecture delivers. The key point is that either way, many people still believe them to be true. And, of course, the incumbent telecom service providers have a vested interest in propagating these myths... er, beliefs.
The NEW architecture, thanks to its key components – WAN Optimization, server virtualization, distributed/replicated file service, colocation, and WAN Virtualization – addresses the underlying issues for those "myths" which actually have some historical validity.
Over the next few columns, we will be debunking these myths, and in the process better arming forward-looking enterprise WAN managers with the information they'll need to sell a NEW architecture-based approach to their WAN – an approach that will allow them to both address the concerns of the past and the present, as well as handle the future reality that the enterprise WAN always needs more bandwidth, and the need to provide a cost-effective, reliable solution to supporting enterprise use of public and hybrid cloud computing.
A twenty-five year data networking veteran, Andy founded Talari Networks, a pioneer in WAN Virtualization technology, and served as its first CEO, and is now vice president of product management at Aryaka Networks. Andy is the author of an upcoming book on Next-generation Enterprise WANs.