Unlike compression appliances, Riverbed delivers more bandwidth along with the rest of the solution—transport and application-level protocol optimization. Only Riverbed delivers a complete solution to the problem of application throughput.

Without accounting for the impact of latency, merely compressing data offers little to no improvement in throughput beyond relieving congestion. WAN compression appliances typically reduce WAN traffic three to five fold (this is also referred to as WAN expansion by some vendors). While some vendors imply that compressing data leads to improved performance, in most cases it does not.

Try thinking of your WAN problems as a traffic jam. If a highway has two lanes in each direction and is completely congested, traffic may only be able to travel at 30 mph instead of 60 mph. By adding two more lanes in each direction (adding bandwidth), or by squeezing everyone’s car into half the size (compression), the congestion can be relieved and the cars can travel the speed limit of 60 mph. In that sense, the performance of the cars (the application) has improved.

However, once you have relieved the congestion through compression or by adding bandwidth, the throughput of the car doesn’t improve with additional capacity. So even if you added 10 more lanes, the cars would still be going only 60 mph (assuming they obeyed the speed limit).

For example, if you were driving a truck on that 12-lane highway and had to deliver 100 truckloads of goods between two cities that were 30 miles apart, the job would take 100 hours, no matter how many additional lanes were available.

The same is true on a WAN. If an application like Windows generates 1000 round trips to accomplish a task, and each trip takes 100 ms, then the task will take 1 min, 40 seconds (100 seconds), not matter how much bandwidth is available. This is why adding compression technologies often has little to no impact on throughput, even though the bandwidth consumption may go down.