Wouldn’t it be nice to get help when you’re in trouble such that you never suffer the effects?
Like the secret service whose job it is to protect the President. They stay out of the way and remain indiscreet and unnoticeable until danger approaches – then they spring into action like a well choreographed army geared to destroy the threat. If they do their jobs right, trouble has been averted and things continue on without any calamity.
When it comes to bandwidth contention, saturation and blackout threats is there any technology whose only mission is to spring into action to avoid this situation? Yes, but as far as I know, only one – Equalizing!
But first, let’s briefly review the many technologies that are marketed and sold under the banner “WAN Optimization” – what do they really do? Most of the time, what they do is try and prioritize one thing over the other – thus accelerating a subset of your applications. In some other cases, they take a XMbs pipe and try to make it feel like a X+YMbs pipe. But what do they do if the majority of applications causing the peak are “high priority,” or what if demand is such that you peak at X+YMbs bandwidth?
In a previous blog article, we wrote about various WAN Optimization technologies and their pro’s and con’s. It’s a great read if you would like more general details about the variety of specific WAN Optimization technologies.
Yet, which of these WAN Optimization technologies only help during peak bandwidth Periods? Do any of them help to manage through those peak periods – by kicking-in to provide relief, and then kicking-out once it’s no longer needed? Isn’t that the time when you need help most and isn’t that the type of help you need?
I get asked the question to describe the differences between the NetEqualizer and other products/technologies all the time. These are good and fair questions, and I like answering them. However, it does demonstrate that product and technology differences are not widely understood by the market place. Not all “WAN Optimization” technologies operate the same way, or solve the same problems.
With each of the non-Equalizing technologies, you can still suffer from peak load problems because, by definition, the other technologies do nothing to dynamically change the bandwidth demand that’s causing the peak load. When using non-Equalizer based technology bandwidth demand is still allowed to increase until the network can’t deal with it anymore. With these technologies the only thing you’ve achieved is extending the point at which the peak may happen, but you’ve done nothing to remove the fundamentals of the peak bandwidth demand. After saturation has occurred, you end up in the same place with the same problems and complaints. That’s not relief and that’s not a solution! More or less, it’s the same thing as buying a little more bandwidth, it may provide some additional breathing room at first, but it really doesn’t solve the problem especially long term.
While most “Wan Optimization” approaches extend or “stretch” bandwidth, ONLY Equalizing dynamically optimizes QoS during peak periods and can “stretch” the effective bandwidth almost indefinitely. To do this, a few applications will be slowed down, but if properly tuned these applications will represent less than 10% of all applications, and frankly, they’re the ones that are the reason for the peak problem anyway – so it’s fair and appropriate.
The fundamental principle of Equalizing is to optimize the WAN by behavior, and only during impacted conditions. An Equalizer will provide all applications the bandwidth they need, until the equalizer senses that there is a peak condition occurring at which time it “kicks” in by slowing down the largest flows while still allowing the smaller flows to pass through the network as a priority and without any delay. When correctly set, large flows should constitute 5-10% of all flows, and by slowing this small percentage of flows to the advantage of all others, the equalizer leverages bandwidth to the benefit of 90-95% of everything else. Since decisions on what to slowdown is done by flow size, and not by port, protocol or content, it is effective on 100% of all flows – including those that are encrypted or otherwise “cloaked.” Most importantly, during peak periods, it will dynamically react by fairly slowing down only the “hogging” applications.
When the peak period ends, the delays are removed, and all traffic resumes in its natural order. This type of WAN optimization is specifically geared towards peak period management, and is highly effective. Furthermore, since it’s driven by a few key parameters – it is a “set-and-forget” product that requires zero ongoing maintenance and doesn’t require any foreknowledge of the types of users, devices or applications that will be confronting the network. That’s a perfect fit for networks that have unpredictable uses like those administered by: ISPs, Schools, Business Centers, Hotels, Libraries, Convention Centers, any public WiFi Network, etc… Wherever you have an unpredictable application load to manage, Equalizing is a far superior method than any other.
It is true peak contention management like no other – and the only technology that helps at the time when you need it most!
Please comment – we’d love to hear your opinions.