Fair Scheduling for Wireless Mesh Networks

1
Fair Scheduling for Wireless Mesh Networks

• Jason Ernst University of Guelph
• Prepared for CS6650 - Mobile Wireless Networks

2
Presentation Outline

• Introduction Background
• WMN, Fair Scheduling, Mobility
• Motivation
• Limitations of Current Implementations
• Related Work Applications
• Current Progress
• FS Implementation in C
• Fair Scheduling algorithm
• Future Work Project Goals
• NS2 Implementation Challenges
• What will be simulated
• Questions

3
Introduction and Background

• Wireless Mesh Network
• Ad-hoc network consisting of mesh routers and
  mesh clients.
• Mesh routers
• have more resources (power, memory, bandwidth)
  than the mesh clients and are used for forwarding
  packets
• Mesh Clients
• A wireless node which may be or may not be
  mobile, may be constrained by resources such as
  battery life

4
Introduction Background
5
What this work tries to address

• Three Assumptions in Hubaux and Salem paper A
  Fair Scheduling for WMN
• Static and known topology of the network
• Nodes are not mobile (both clients and mesh
  routers)
• Nodes cannot be added or removed
• One gateway
• Huge bottleneck point, both for traffic and
  scheduling
• Mesh Routers are not mobile
• Mobile mesh routers allow for interesting
  applications such as mobile transit networks and
  military applications

6
Motivation

• Wireless Mesh Networks
• Fewer gateways required compared to single-hop
  APs spread around
• Redundant paths in case of congestion, failure
• If properly designed can support self properties
  of autonomic networks
• Fair Scheduling
• Ensure every user gets equal service for equal
  money, control greedy or malicious nodes, prevent
  starvation
• Mobile WMN
• Moving nodes can be grouped with a mobile MR to
  minimize the number of handoffs since they are
  all moving together (ex bus scenario)
• A type of clustering where moving nodes are
  paired with a moving MR

7
Related Work Applications

• Applications
• WMN transit system
• Mobile military applications
• Community Mesh Networks
• Last-Hop solutions for ISPs
• Related Work
• Wireless Local Area Networks (WLAN)
• Mobile Ad-hoc Networks (MANET)
• Operating System Scheduling
• Distributed Computing (SHARCNET)

8
Current Progress

• Implementation of A Fair Scheduling for WMN by
  Hubaux and Salem in c
• Uses concept of compatibility matrix to
  produced a collision-free STDMA scheduling for
  the network
• In their paper they do not allow for any mobility
  (mesh clients or mesh routers)
• They do not provide a mechanism for distributing
  the scheduling to the MRs

9
Compatibility Matrix
GW
3
0
MR
MR
1
2
4
5
MR
MR
MR
MR
0 1 2 3 4 5 0 1 0 0 0 1 1 1 0 1 0 1 1 1 2 0 0 1 1
1 1 3 0 1 1 1 0 0 4 1 1 1 0 1 0 5 1 1 1 0 0 1
10
Fair Scheduling

• The Fair Scheduling is determined using the
  compatibility matrix and a concept of gain
  where the links are weighted based on the number
  of clients using the link
• The gain is calculated by the sum of all the
  gains minus the max gain for each group of links
• The group with the maximal gain which does not
  intersect with existing selected groups of links
  is chosen to be added to the scheduling

11
Fair Scheduling
Figure 1 Example STDMA Scheduling taken from A
fair Scheduling for WMN by Hubax and Salem
12
NS2 Implementations Challenges

• Current MAC implementations in NS2
• Many make use of RTS/CTS mechanism for collision
  avoidance (802_11) and are complicated
• Some of the more basic TDMA scheduling MAC
  implementations are meant for single-hop
• Some MAC implementations dont work with existing
  routing protocols such AODV (WiMax module)

13
What will be Simulated

• Small to medium sized network (up to about 50
  nodes)
• Increasing number of mobile mesh routers,
  performance will be evaluated with an increasing
  number of mobile mesh routers
• Similarly, increasing number of gateways to see
  the performance changes
• Dynamic network topology (adding, removing and
  moving nodes)

14
Future Work Project Goals

• Continue to work on distribution method to get
  the scheduling to the MRs in the network
• Either centralized via gateway control (as in the
  Hubaux and Salem paper) or a more distributed
  approach where the matrix is split up at each
  layer of the hierarchy
• Implement new features
• Support for multiple gateways (requires
  distributed solution if GW controls the
  scheduling
• Add mobility support for MRs
• Allow dynamic topology changes

15
References

• Agrawal et Al. Achieving Load Balancing in
  Wireless Mesh Networks Through Mulitple Gateways.
  IEEE. 2006. 807-812.
• Bejerano, Yigal., Han, S-J., Kumar, Amit.
  Efficient Load-Balancing Routing for Wireless
  Mesh Networks. 2007. Computer Networks. 51.
  2450-2466.
• Chandranmenon et. Al. On the Design and
  Implementation of Infrastructure Mesh Networks.
  IEEE Workshop on Wireless Mesh Networks (WiMesh)
  2005.
• Cheng, S-M., Lin, Phone., Huang, Di-Wei., Yang,
  Shun-Ren. A Study on Distributed / Centralized
  Scheduling for Wireless Mesh Network. 2006. IWCMC
  06. ACM. 599-604.
• Gupta, Piyush., Sankarasubramaniam, Yogesh.,
  Stolyar, Alexander. Random-Access Scheduling with
  Service Differentiation in Wireless Networks.
  2005. IEEE. 1815-1825.
• Erwu, Liu., Shan, Jin., Gang, Shen., Luoning,
  Gui. Fair Scheduling in Wireless Multi-Hop
  Self-Backhaul Networks. IEEE AICT/ICIW 2006.
• Hubaux, J-P., Salem, Ben Naouel. A Fair
  Scheduling for Wireless Mesh Networks. WIMESH.
  2005
• Koutsonikolas, Dimitrios., M. Das., Saumitra.,
  Hu, Charlie, Y. An Interference-aware Fair
  Scheduling for Multi-cast in Wireless Mesh
  Networks. 2008. Journal of Parallel and
  Distributed Computing. 68. 372-286.
• Popa, Lucian., Rostamizadeh, Afshin., Karp,
  Richard, M., Papadimitriou, Christos., Stoica,
  Ion. Balancing Traffic Load in Wireless Networks
  with Curveball Routing. 2007. Mobihoc 07. ACM.
  170 179.

16
References contd

• J. Thomas, Cross-Layer Scheduling and Routing
  For Unstructured And Quasi-Structured Wireless
  Networks
• M.S. Kuran, G. Gur, T. Tugcu, F. Alagoz,
  Cross-Layer Routing-Scheduling in IEEE 802.16
  Mesh Networks, in Mobilware08. Austria, 2008.
• M. Neely, R. Urgaonkar, Cross-layer adaptive
  control for wireless mesh networks, in Ad Hoc
  Networks Vol . 5, pp 719-743, 2007.
• J. Tang, G. Xue, W. Zhang, Cross-Layer Design
  for End-To-End Throughput and Fairness
  Enhancement in Multi-Channel Wireless Mesh
  Networks, in IEEE Transactions on Wireless
  Communications, Vol. 6. pp 3482-3486. October
  2007.
• X. Wang, K. Kar, Cross-Layer Rate Control for
  End-to-End Proportional Fairness in Wireless
  Networks with Random-Access, in MobiHoc 05.
  Illinois, USA, 2005.
• J. Tang, G. Xue, C. Chandler, W. Zheng, Link
  Scheduling with Power Control for Throughput
  Enhancement in Multihop Wireless Network
• V. Kawadia, P.R. Kumar, A Cautionary Perspective
  on Cross-Layer Design, in IEEE Wireless
  Communications, pp 3-11, February 2005.
• I.F. Akyildiz, X. Wang, Cross-Layer Design in
  Wireless Mesh Networks, in IEEE Transactions on
  Vehicular Technology, Vol. 57, 2, pp 1061- 1076,
  March 2007.

17
Questions Comments?

• Jason Ernst
• jernst_at_uoguelph.ca
• University of Guelph