Will WiFox fix Wi-Fi?
22nd Nov 2012 | 16:22
Software could be a simple solution to poor connections
For anyone who spends a significant amount of time online, it was hard not to get excited when news of WiFox and its potential to massively bolster Wi-Fi speeds emerged last week.
The prospect of having our internet connections given a 700 per cent boost should appeal to all and, although it may sound a little hard to believe, the tech whizzes behind the software insist it works - and it won't cost a bomb.
"WiFox observes the number of backlogged packets at a Wi-Fi access point," explained PhD student Arpit Gupta.
"Based on backlog size, it decides every 100 milliseconds whether to prioritise access point traffic over other users' data traffic. It then uses its mapping algorithm to decide the level of prioritisation on scale of 10."
Gupta is just one of the brains behind WiFox, a project spearheaded by North Carolina State University's Computer Sciences lab. The team is already receiving proposals from various service providers around the globe to share the technology with.
Gupta explained that their ultimate goal was to ensure WiFox reaches everyday users, saying that the lab "focuses only on projects that can make a real impact".
The easiest way to explain WiFox is by thinking of it as a traffic cop. By analysing the amount of data travelling through different Wi-Fi channels, the software is able to clear up congested channels and divert traffic elsewhere, allowing data to flow freely again.
By prioritising traffic, WiFox enables each connected user to send and receive data at a more equal rate, meaning less chance of users hogging bandwidth with nightly Call of Duty marathons or endless movie streaming.
But WiFox could prove something of a double-edged sword for some. Because the software is essentially a traffic management system, it could end up a hindrance to those who partake in bandwidth-heavy activities like online gaming, essentially slowing down their connections to allow for more equal traffic flow across the network.
But because it's early days for WiFox, it's not clear yet how it will handle this or if it's even an issue at all.
"Wi-Fi's core premise is pretty simple," said Adrian Edwards, consumer solutions manager at D-Link UK and Ireland.
"Routers and adapters send and receive data using radio waves. The two primary wireless frequencies (2.4GHz and 5GHz) are divided into channels of which your router can use one or two simultaneously. The idea is that by using different channels, neighbouring networks avoid stepping on each other's toes.
"However, with the rapid growth of wireless networks and technologies it is becoming a lot easier for channels to become congested. WiFox is an interesting proposition [which] could go a long way to helping relieving channel congestion."
WiFox's ingenuity lies in its practicality, in that massive and costly overhauls of already-existing networks won't be necessary in order to reap the software's benefits. In fact, integrating WiFox into the public network should be really, really easy.
"The most important thing is that you do not need to change existing hardware for WiFox - it can be rolled out as a small firmware update," explained Gupta.
"What this implies is that small public Wi-Fi service providers can test our scheme over their own test networks and roll it out as a software update. It will not require any update for users' devices, [meaning] rolling it out on a larger scale will not be a complex task."
In tests, Gupta and his team managed to improve WiFi performance from between 400 and 700 per cent - an encouraging figure.
But these were exactly that: tests. Surely the same results couldn't be achieved in a public hotspot where connected users can reach into the hundreds? Gupta is confident they can.
"When developers and other vendors further test WiFox on different network scenarios, they may not get the exact same results as ours, but we expect them to be quite close to what we observed," he said.
Amazingly, WiFox's performance actually improves depending on the weight on the access point, meaning the more people trying to connect at once, the quicker it clears up data channels.
At its core, the idea behind WiFox appears relatively simple. In which case, why hasn't anything like it been put forward already?
"When we started working on this problem, there were few solutions already proposed on this topic," explained Gupta.
"We realised that none of them could actually be achieved in a practical manner as they either required hardware changes, changes to existing MAC protocols or required changes to network stack for both WiFi access points and associated users. We started from the top down, figuring out how the final solution should look like and then worked further to make it happen."
Gupta will be presenting a paper proposing the tech - of which he is co-author - at the ACM CoNEXT 2012 conference in December. From there, the team's aim is to make WiFox commercially available as soon as viably possible. That said, they still have some way to go.
"We are trying to figure out what should be our exact strategy to ensure mass usage," said Gupta. "We expect that after our talk at CoNEXT we can draw more serious attention from industries for real network testing and deployment.
"We're currently in talks with various vendors to ensure its commercial delivery, but everything is still in a very early phase."