Adding Bandwidth Alone Does Not Solve the Problem
Over the past few years, with dropping network prices, it was tempting to add bandwidth
to try and solve application problems on the WAN. But bandwidth is often not the
answer. The problem is usually a bit more complicated … it is an issue of physics.
Consider a completely clear 64 KB link traversing a distance of 10,000 kilometers
(about the distance between Tokyo and New York). Under the best possible lab environment
conditions (light traveling in a vacuum), a 1 KB packet would take a little more
than 33 milliseconds to make the trip. What if the links' bandwidth is raised to
155 MB … how long would the same packet take to make the trip? Exactly the same
33 milliseconds. Why is this? The problem is not bandwidth, it is latency. The packet
cannot travel faster than the speed of light. Simply adding more bandwidth will
not improve the latency on the WAN link.
The reality is that average latency between Tokyo and New York is closer to 170
milliseconds … almost six times worse (real-world versus lab environments). Even
so, a single 170 millisecond delay on application response may not sound so bad.
Unfortunately, that is not the whole story. Most of the applications that run the
business are fairly chatty. The number of roundtrips necessary to download a typical
home page is in the hundreds, if not thousands. In the end, distance will contribute
to latency, and even small increases in latency can dramatically impact the performance
of applications. Adding more bandwidth will not improve performance in a significant
way.
Mitigate the Effects of Latency
Optimization: TCP Acceleration
With the current application technology, there is no way to increase the speed of
light. This means there is no way to really reduce latency when distance is the
issue. However, the Expand Acceleration solution is able to mitigate the effects
of latency. This is accomplished via a deep understanding of the protocols applications
use to communicate. Essentially, with this understanding, it is possible to implement
techniques that will change the nature of TCP communication to a latency mitigating
environment. Add this to Expand’s Virtual Bandwidth techniques, and the user experience
improves substantially. Together, these optimizations will speed the response time
of applications, directly improving the productivity of business.
Controlling Variable Environments
Optimization: Dynamic Congestion Control
While methods to mitigate WAN latency are extremely powerful, Wide Area Networks
are variable environments. Congestion, lossy environments, and variable bandwidth
conditions can significantly impact application delivery. Expand has the industries
most advanced congestion control technologies. By monitoring WAN links on a packet
by packet basis in real-time Expand can use network information to actively adjust
its network conditioning technology. Expand’s solution literally adjusts itself
to provide the very best optimization for network conditions, in real-time. Using
Dynamic Congestion Control and real-time Variable Bandwidth Adjustment technologies
Expand’s solution stays on top of the ever-changing WAN challenges, actively conditioning
your link to serve your business best.
Average Virtual Bandwidth Gain
Average Expected User Performance Gain
