A Proximity-Aware Transparent Handoff Mobility Scheme for VoWMN (PATH)
Last year, I dedicated an important part of my work and time at the research center of my University to develop a way to provide fast handoffs within a 802.11s network. This post briefly describes it, named PATH after several attempts to make it sound cool.
Most of the inspiration came from a great and interesting project called SMesh, a fast handoff mobility system created by the DSN Labs staff at Johns Hopkins University. One of the people behind the system is @ralucam, who helped me to understand the protocols of this system.
Yet another work that was taken into consideration is LCMIM, which is a Light-weight Client Mobility approach for Infrastructure Mesh networks. It is basically a simpler solution than SMesh but it is somewhat inefficient as it continuosly contends for the medium by broadcasting gARP messages, flooding the channel even though no Client is really using the network and that is bound to be used with a reactive routing protocol, in the case presented in the paper, AODV. The good side is that its nodes maintain independence and that it’s a simpler solution as well.
The PATH scheme
After several months of hard work and going through a learning process, I developed PATH, a Proximity-Aware Transparent Handoff mobility scheme. I should state that it is by no means a solution as complex and complete as SMesh as it only addresses the reversed channel latency issues caused by handoffs, but it certainly is simpler to install, follows a simpler logic and is independent of the type of routing protocol used as well.
It all went good. It basically sends gARP messages based on client proximity (which is measured by the perceived RSSI on each node), thus associating clients to new nodes and switching its connectivity. Take a read to the presentation!. It should be noted that all three: SMesh, LCMIM and PATH create an infrastructure Mesh Network by setting wireless nodes into adhoc mode. By using this mode, all three schemes do not let the client decide when to roam but leave the responsibility to the network nodes.
It was designed to provide fast handoffs for real-time voice applications. Although it is essentially a work in progress and there are lots of ways to improve the scheme, I had the opportunity to share the results and they were all great🙂
I will be posting a link to the paper soon!. Hope you like the presentation.