Editor's note: More complex networks and applications mean that enterprises must design WANs capable of handling...
ever-increasing amounts of traffic without sacrificing performance. Read the first article in our three-part series on buying WAN optimization tools to help you make the right investment. In part 2, check out WAN expert Kevin Tolly's advice on how to benchmark WAN optimization tools before you buy. Finally, in part three below, find out what steps you should take to make sure your WAN optimization appliance or platform to suit your needs.
After you buy the right tools, WAN optimization configuration options will enable you to instruct your optimization platform to map the relative importance of your traffic streams. As you review your choices, pay special attention to the ones that allow you to specify the relative priority of one application stream (usually transaction traffic) over another (usually file transfer or other batch operations.)
Remember, you want the test traffic to resemble your actual traffic as closely as possible. Most importantly, aim to generate multiple user sessions. Many WAN optimizer platforms will prioritize on a session basis. As a result, sending a large amount of traffic over a single session might not allow the WAN optimization system to operate properly. Again, the goal is to make the test traffic -- the application type, content and number of different sessions -- match your actual traffic as closely as possible.
Scripting tools can help WAN optimization
As you review your choices, pay special attention to the ones that allow you to specify the relative priority of one application stream.
To provide a repeatable test scenario, it is useful to use scripting tools that interact with user screens to run the test. These can be anything from freeware AutoIT scripts to more complex capabilities provided by commercial products like Login VSI for VDI environments.
Now all that's left to do is to run the test and assess the results. While a WAN optimization platform might provide you with some statistics, the real measure of effectiveness is application performance as viewed by the user.
Fortunately, you can just focus on the performance of a single, "key" user session to gauge the impact of WAN optimization on your important traffic. This can be done in three basic steps.
First, run the key application across the simulated WAN with no other traffic and with the WAN platform disabled. Measure the response time, either by individual interaction or by beginning to end, depending upon what is important to you.
Next, with WAN optimization still disabled, start all your other traffic, then your key application. Measure the response as you did previously. (Hint: If it didn't get worse, either you aren't sending sufficient traffic to burden the WAN or you don't need WAN optimization after all.) Provided you produced the expected degradation in user performance in the previous step, it's time to let WAN optimization make things better.
Making sure WAN optimization is working correctly
To do that, enable WAN optimization, start your background traffic, then run your key application and measure the response time. You should see your key application performing better than it did in the previous run when WAN optimization wasn't active.
More on buying and tuning up a WAN optimizer
Buying WAN optimization tools: What you need to know
How to benchmark WAN optimization tools before you buy
You can also measure effectiveness by measuring the amount of traffic traveling across the WAN to the headquarters site, then measuring the amount of traffic WAN optimization delivers to the target LAN. The higher the ratio of LAN traffic to WAN traffic, the more the WAN optimization is squeezing out of your WAN. For example, if a WAN optimization deployment can deliver 8 Mbps traffic to the target LAN from your 2 Mbps WAN, that indicates it's producing a better throughput and likely a better end-user response time than a WAN optimization deployment that can deliver only 4 Mbps of traffic in the same environment.
While it is a little bit of work, only by benchmarking your applications in the lab can you have any confidence in the benefit that WAN optimization can bring to your production environment.