Mar 11, 2015
In one of my last two blogs I discussed how a software defined WAN will automate the configuration of WAN edge routers and enable a WAN with multiple links to automatically reallocate traffic based on changing network conditions. In the other I discussed some of the WAN design options that the Open Networking User Group (ONUG) have identified. Based on just those two blogs an enterprise architect might well ask: Is that all there is to WAN design – dynamic load balancing over MPLS and the Internet?
My thinking on network design is heavily influenced by a scene in Alice in Wonderland that is comprised of the following dialogue between Alice and the Cheshire Cat:
Alice: “Would you tell me, please, which way I ought to go from here?”
Cheshire Cat: ‘That depends a good deal on where you want to get to.”
Alice: “I don’t much care where.”
Cheshire Cat: “Then it doesn’t matter which way you go.”
To paraphrase the preceding dialogue, if you don’t know how your company will use its WAN, then just choose any design. Looking at that conversation in the context of the title of this blog: in addition to implementing dynamic load balancing, a key characteristic of a next generation WAN is that it implements functionality that addresses the specific challenges associated with how a company will use their WAN.
The 2014 State of the WAN Report contained market research that showed that, by a fairly wide margin, the biggest driver of increased Internet traffic is providing connectivity to public cloud computing services. Unfortunately the latency of the Internet tends to be higher than the latency associated with an enterprise WAN service, such as MPLS. One of the reasons for this is the use of the Border Gateway Protocol (BGP) to route the traffic that goes between the millions of networks that comprise the Internet. Since BGP doesn’t know the specific, real-time characteristics of the networks that comprise the Internet, the end-to-end path chosen by BGP isn’t likely to be the path with the least latency.
To illustrate how to overcome this challenge, assume that your company wants to make use of a SaaS service like Office 365. In virtually all instances, your company will not be able to place any optimization or security functionality on the SaaS provider’s premises. You can, however, leverage functionality that is placed in a gateway as close as possible to the SaaS provider’s premises. That functionality could be deployed in one of your company’s facilities or hosted at a communications service provider’s POP, within a cloud hub (i.e., Equinix, etc.) or at an IaaS provider’s facility (i.e, AWS, Azure).
Unlike the current environment that relies on BGP, functionality that understands the real-time characteristics of the Internet, such as loss and latency, between your users and each potential gateway of a SaaS service can be leveraged to improve performance. This information should complement software strategically placed at each gateway in order to not only perform the optimal routing, but to also distribute information about adverse Internet conditions to other software instances across the network.
In addition to adverse Internet conditions, another challenge associated with providing access to a SaaS services is that new services are introduced frequently and service components and data centers are regularly added to existing SaaS services. As a result, your company needs cloud intelligence to aggregate information about the changing subnets and IP addresses for SaaS services so that the software functions that perform the advanced routing have the most current information about those services.
Over the last decade there hasn’t been a fundamentally new transmission service introduced into the market. An understandable result of the absence of new transmission services is that most network organizations have put WAN design on cruise control. The emergence of a software defined WAN complimented by a wide array of optimization, security and management technologies indicates that it’s time for network organizations to take WAN design off cruise control and to analyze the growing set of WAN design options.
Image credit: Wikimedia Commons
Jim has a broad background in the IT industry. This includes serving as a software engineer, an engineering manager for high-speed data services for a major network service provider, a product manager for network hardware, a network manager at two Fortune 500 companies, and the principal of a consulting organization. In addition, Jim has created software tools for designing customer networks for a major network service provider and directed and performed market research at a major industry analyst firm. Jim’s current interests include both cloud networking and application and service delivery. Jim has a Ph.D. in Mathematics from Boston University.