When you think Wide Area Network, “WAN,” and you compare it to your local LAN, you think of the terms of a “high cost,” “low bandwidth” and “high latency” connection to an ISP. Whether the link is a T1 or T3 circuit or a fiber SONET-based connection (OC-3, OC-48), bandwidth demand inevitably outstrips bandwidth supply.

By the way, "WAN" is really just a generic term as vague as the word "network." It doesn't really mean any one thing, but has become the acronym for discussing non-local network connections.

WAN connectivity develops problems because business-critical applications generally stay in direct competition with all other Internet traffic on your link. Unwanted traffic can, and often does delay business applications. Spam, malware and worker-invited Web traffic like peer-to-peer, Skype, You Tube and others can at times hinder a business's ability to complete its mission.

The real culprit is TCP. It really doesn't care about anything but getting things through reliably. Unless some kind of engineering has been done to prevent it, all applications get treated equally. And TCP will cheerfully use all available bandwidth in a very bursty, inconsistent manner.

TCP's only congestion control mechanism is to simply start sending traffic slowly, and then increase speed until loss occurs. When loss happens, the lost data must be retransmitted. That creates even more congestion.

Curbing Congestion
VistaOne can place optimizers at the entrance to each network that will conduct site-to-site optimization. Specified TCP connections are then terminated locally and the WAN-facing link transmits a proprietary TCP/IP protocol optimized for that link. Simply taking TCP out of the picture in this situation will dramatically improve throughput.

Of course, that's just for site-to-site business applications with two configured endpoints. If the WAN link is completely saturated with other Web traffic, the modified streams won't improve things much.

Solutions can involve quality of service (QoS) that identifies which traffic is critical and which isn't. This has traditionally been accomplished by marking or coloring packets. Class then determines which traffic gets priority.

Again, we're in a situation where we’ve prioritized important traffic but still have a congested link. We can implement some type of queuing or traffic shaping mechanism as a workaround. This technique classifies all traffic into queues. The most important data gets processed first, and the rest later. When congestion already exists on a link, all queuing accomplishes is to slow everything down but the high-priority traffic.

When we realize that the level of optimization required very much depends on the specific application, we begin to formulate the ideal solution. Prioritize traffic and guarantee a certain amount of available bandwidth for mission critical applications. Identifying urgently critical, high availability traffic like ERP and CRM applications and placing a higher priority on that traffic than we do on Web browsing or video goes much farther toward ensuring efficiency.

Complete WAN optimization solutions allow a business to do much more than simply queue the bad traffic. They can block unwanted (in and outbound) traffic, permit it at certain time during the day, give priority to specific hosts and enforce many other related policies. They will optimize the actual traffic as well, providing lower latency and higher throughput for the most critical applications.

Broadly speaking, all WAN optimization solutions boil down to a few available techniques:

• Traffic prioritization
• End-to-end tunnels (employing better protocols than TCP)
• TCP tricks (selective acknowledgements or ACKs, limiting retransmissions, reordering packets, and compression)