Research Idea

Lei RaoMarch 7th, 2009

Related Work

[1] J.G. Dai and W. Lin, “Maximum Pressure Policies in Stochastic Processing Networks,” Oper. Res., vol. 53, 2005, pp. 197-218.

[2] S. Kandula, K.C. Lin, T. Badirkhanli, and D. Katabi, “FatVAP: aggregating AP backhaul capacity to maximize throughput,” Proceedings of the 5th USENIX Symposium on Networked Systems Design and Implementation, San Francisco, California: USENIX Association, 2008, pp. 89-104.

Problem 1: AP Utilization in WiFi Networks

1 2 3 4 5

1 2 3 4 5

1 2

Five Laptops

Five Virtual Buffers

Two APs

Problem: How to assign the utilization time of each AP to each laptops so as to obtain maximum throughput in WiFi networks? Note that APs and Laptops are one-to-one mapping.

Solution: Maximum Pressure Policy

Assumptions:

• Buffers are never empty, i.e., there are always data in laptops to transmit.

• APs are always busy, i.e., at one time instant, each AP is connect to one laptop.

Decision Variables:

• Xij: long-run time fraction of AP j assigned to laptop i.

Definition of Maximum Pressure:

• The maximum packet quantity in the utilized buffers under a certain allocation of APs.

Features of the solution:

• Maximum Throughput

• Minimum Average Delay

Problem 2: Dynamic Routing in Real Time Sensor Networks

1

2 3

4 5

6 7 8

9 10

s1

s2

d1 d2

Route 1 for s1

Route 2 for s1

Route 1 for s2

Route 2 for s2

Route 3 for s2

For Each Node: (Node 3, 4, 5 as an example)

1 32

1

32

Input Processor

Virtual Buffer

Service Processor

Solution: Maximum Pressure Policy

Advantages:

• Maximum Throughput Guarantee

• Instant Routing Decision (Dynamic)

• Minimum Average Delay

Disadvantages:

• Not Purely Local (Local and Downstream Information)

• Not Hard Real Time Guarantee