Wireless MESH Networks: PHY/MAC Optimization

Principal Investigator
Sumit Roy

Sponsor(s)
Intel, Corp.

Award Period
08/29/2005 - 08/28/2008

Abstract
A suitable architecture for a 802.11 based Broadband Wireless Access Network (BWAN) that provide client devices with access to a backbone network (the Internet) while meeting next generation expectations continues to be an issue of considerable interest. A possible architecture of such networks will consist of two tiers – a bottom tier of mobile clients who are associated with a fixed second tier of nodes that are inter-connected wirelessly to provide broadband access to the end-users. The second tier forms a MESH (i.e. a multi-hop network) comprising of two types of nodes: a few wired gateway nodes/APs with a back-haul connection to the Internet and a predominant set of (cheaper) wireless router APs (no back-haul connection) whose primary function is to forward packets for to/from other routers. Such an architecture has several compelling advantages - making this work with predominantly cheap `soft’ AP routers underscores its potential for cost-effective network scaling. Further, the architecture can exploit the advantages of adaptive mesh network concepts to provide efficient and robust multipath routing.

However, many of the key protocol stack design/engineering issues for such mesh networks remain unanswered at this point. Clearly, a primary attribute of any BWAN design must be network scalability - since the BWAN is effectively an intermediate transport mechanism between the end-devices and the backbone network, it’s throughput must scale with a) increasing peak data rates per client and b) larger number of clients.

Clearly, allowing multiple simultaneous transmissions while suitably managing mutual interference will be key to aggregate scaling. However, emphasis on scaling aggregate throughput needs to be balanced with the need for notional fairness between users (which is very MAC and network topology dependant).

Thus, design of a successful MESH network must fundamentally explore cross-layer approaches to optimization of the protocol stack, in particular, tuning parameters jointly at layers 1, 2 for end-2-end network optimization. Specifically, our research agenda will focus on analysis and simulation of .11 MESH networks to uncover principles for:

• Joint PHY/MAC optimization for enhanced aggregate 1-hop MAC throughput;

• Joint evaluation of link layer models, channel allocation and routing metric on end2end network performance

Updates or corrections to this page should be sent to gheaton@u.washington.edu.