Traffic Balancing in Wireless MESH Networks
| Xiaojing Tao, Thomas Kunz, David Falconer | |
| Systems and Computer Engineering | |
| Carleton University | |
| tkunz@sce.carleton.ca |
Traffic in Wireless MESH:
Potentially Unbalanced
| Explained in previous presentation for MANETs | |
| Nodes measure medium busy time, decide whether node is in congested area based on dynamically determined threshold | |
| Sender picks route to destination (here: one of the gateways) that has lowest number of overloaded nodes | |
| To route towards a node in a MESH: have to trigger router discovery from ALL gateways (or have destination trigger “reverse” route discovery to all gateways) |
Simulation Results I:
2 Gateways, high mobility
Simulation Results
II:
2 Gateways, low mobility
Interesting Issue: Gateway/AP Location
Performance vs. Gateway Location
| Mesh networks: Gateways/APs are performance bottlenecks | ||
| MANET routing protocols, routing to closest AP, tend to (temporarily) overload some APs | ||
| TB can improve network performance by distributing traffic more equally to all APs | ||
| Future Work: | ||
| Location of APs has major impact on performance | ||
| IEEE 802.11 is poor protocol for APs, as they need to access media more frequently than nodes | ||
| MESH are more static, should we re-issue route discovery periodically? | ||
| Explore TB in our wireless mesh testbed, based on IEEE 802.11 a/g radios and Intel IXP 425-based wireless routers | ||