The authors propose an architecture for wireless mesh network that has four goals:
- Unconstrained node placement
- Omni-directional antennas
- Multi-hop routing
- Optimization of routing for throughput in a slowly changing network
These goals provide the ease of deployment which is an important requirement for community wireless networks. The evaluation of the architecture is done by deploying 37 nodes spread over 4 square kilometers of urban area.
The evaluation determines the end-to-end performance of Roofnet and the effect of the design decisions. The measurements are based on:
The evaluation determines the end-to-end performance of Roofnet and the effect of the design decisions. The measurements are based on:
- Multi-hop TCP data: 15 second one-way bulk TCP transfer between each pair of Roofnet nodes
- Single-hop TCP: direct radio link between each pair of routes
- Loss matrix: loss rate between each pair of nodes using 1500-byte broadcasts
- Multi-hop density: TCP throughput between a fixed set of four nodes varying the number of Roofnet nodes that are participating in routing
The end-to-end performance of Roofnet is determined by the basic performance, link quality and distance, effect of density and mesh robustness.
I liked the fact that the paper uses a network that is deployed in the real world to evaluate the architecture. This aids in understanding the real-time traffic and how the network works in such scenarios. Another thing to take away from the paper is how the authors have taken great care to provide ease of deployment and maintenance. This aided in deployment of Roofnet by volunteer users.
The evaluation shows that the average throughput decreases to 43% if the best two nodes are lost. This can be a huge problem because you are relying on these two nodes for higher throughput. Another thing that is not clear to me is how crucial is their placement of few nodes at the top of tall buildings? The paper lacks investigating the architecture in terms of the dynamics of the network. It does not handle the effects of user joining/departing and also about varying network performance over time.
I liked the fact that the paper uses a network that is deployed in the real world to evaluate the architecture. This aids in understanding the real-time traffic and how the network works in such scenarios. Another thing to take away from the paper is how the authors have taken great care to provide ease of deployment and maintenance. This aided in deployment of Roofnet by volunteer users.
The evaluation shows that the average throughput decreases to 43% if the best two nodes are lost. This can be a huge problem because you are relying on these two nodes for higher throughput. Another thing that is not clear to me is how crucial is their placement of few nodes at the top of tall buildings? The paper lacks investigating the architecture in terms of the dynamics of the network. It does not handle the effects of user joining/departing and also about varying network performance over time.
3 comments:
Nice summary of the design goals and evaluation.
In fact, I referred to in my summary to your summary and wrote about details of mechanisms I found interesting.
I was also concerned about the loss of two nodes causing a 43% decrease in throughput. Good analysis of the missing evaluations, I didn't think of those!
Post a Comment