Interactive WiFi Connectivity for Moving Vehicles

Summary

This paper presents an evaluation of 6 different handoff protocols for WiFi in moving vehicles. It also proposes ViFi, a protocol that exploits basestation diversity to minimize disruptions and support interactive applications for mobile clients.

For the evaluation, they test these protocols in a 2-month long experiments on 2 platforms. One is VanLAN, which consists of 11 basestations and 2 vehicles used for shuttle service around the Microsoft campus in Redmond. The other consists of vehicles in Amherst. One of these vehicles logs all the beacons heard from all the surrounding BSes.

They test the following 2 applications: VoIP and short TCP transfers.

When evaluating the different handoff protocols, they use 2 metrics:
1. aggregate performance: considers total number of packets delivered and the total time or distance over which the vehicle is connected.
2. periods of uninterrupted connectivity: measures contiguous time intervals when the performance of the application is above a certain threshold.

The 6 handoff policies evaluated are:
1. RSSI
2. BRR
3. Sticky
4. History
5. BestBS
6. AllBSes

When testing performance related to BS density, they found that all protocols' performance increases with density but that relative performance of protocols stays the same throughout. And, AllBSes is found to have the best performance while History, RSSI and BRR are found to perform similar in most cases.

When looking at uninterrupted connectivity results, they found that AllBSes had the fewest regions of inadequate connectivity (because it uses multiple BSes), while BestBS had a few more and BRR had the most.

Then, the authors explain the design and implementation of ViFi. ViFi, motivated partially by AllBSes, leverages BS diversity to reduce disruptions. ViFi designates one of the BSes as the anchor, which is responsible for connecting to the internet. The other BSes are designated as auxiliary.

The authors evaluate ViFi and find that the link-layer performance of ViFi is close to ideal. ViFi has a two-fold application performance improvement over current handoff methods.

Criticism & Questions

I liked reading this paper. It did a great job of evaluating the current protocols and then building off one of those protocols to come up with ViFi. They also had a solid implementation and evaluation of ViFi.

What I am most curious about is the practical application of WiFi connectivity in moving vehicles. They motivate their work by mentioning the growing need and opportunity. I would have liked them to expand more on that and possible current use cases of this technology.

In addition, I would like to know more about what the topology of cities that are covered by WiFi are. When I think about the possible uses for interactive WiFi, I would imagine it would be most useful in longer trips, which generally means crossing multiple cities. If wifi is unavailable in many the cities along the path, that would impact the practical application of this technology.