1 / 27

18 th International Conference on Software Telecommunications and Computer Networks

IEEE SoftCOM 2010 Bol (Brac), 23-25 September 2010. MPR-based Pruning Techniques for Shortest Path Tree Computation. Juan Antonio Cordero Équipe Hipercom -- INRIA Saclay (France). 18 th International Conference on Software Telecommunications and Computer Networks.

drew
Télécharger la présentation

18 th International Conference on Software Telecommunications and Computer Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. IEEE SoftCOM 2010 Bol (Brac), 23-25 September 2010 MPR-based Pruning Techniques for Shortest Path Tree Computation Juan Antonio Cordero Équipe Hipercom -- INRIA Saclay (France) 18th International Conference on Software Telecommunications and Computer Networks

  2. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • Motivation Agenda • Motivation • Topology Pruning in Link State Routing • MPR for Topology Pruning • Path MPR • Impact of the Path MPR Modification

  3. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Motivation (1) Flooding in Wireless Ad hoc Networks • Pure flooding • Redundant retransmissions • Collisions • Reduction of effective bandwidth • Multi-Point Relaying (MPR) • Only a subset of the neighbors are allowed to relay (forward) a message

  4. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Motivation (2) Multi-Point Relays (MPR) MPR coverage criterion Every 2-hop neighbor of the source is covered by (at least) one relay

  5. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • Topology Pruning in Link State Routing Agenda • Motivation • Topology Pruning in Link State Routing • MPR for Topology Pruning • Path MPR • Impact of the Path MPR Modification

  6. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Topology Advertisement 1 Topology Advertisement 2 Topology Acquisition Shortest Path Tree Computation . . . Topology Advertisement n Topology Pruning Topology Pruning in Link State Routing • Link State Routing • Wireless ad hoc networks  Scarce and shared BW • Not all links are necessary for Shortest Paths computation

  7. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • MPR for Topology Pruning Agenda • Motivation • Topology Pruning in Link State Routing • MPR for Topology Pruning • Path MPR • Impact of the Path MPR Modification

  8. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Topology Advertisement 1 Topology Advertisement 2 Topology Acquisition Shortest Path Tree Computation . . . Resulting overlay Topology Advertisement n MPR for Topology Pruning (1) Requirements Rest of the network Computing node • Requirements for the resulting overlay: • Connection • Preservation of Shortest Paths

  9. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 7 1 7 1 1 2 3 2 3 6 6 4 5 4 5 8 8 MPR for Topology Pruning (2) Overlay Connection • The overlay of links connecting nodes to their MPRs is NOT necessarily connected. • Every connected component is dense. • The overlay of links connecting nodes to their MPRs together with all the links of ANY node in the network IS ALWAYS connected.

  10. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 MPR for Topology Pruning (3) Preservation of Shortest Paths (from a node s) The local shortest path between two nodes x and y, reachable in 2 hops, is the path in 2 or less hops between x and y with minimal cost. local reverse shortest path to x network-wide direct shortest path from s 4 2 network 1 min.cost 3 1 x s 2 7 2 For an overlay that includes all links from s, Preservation of links to 1-hop neighbors providing local reverse shortest paths to x, x  Preservation of network-wide direct shortest paths from s

  11. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 MPR for Topology Pruning (and 4) Summary of Results • Overlay of MPR links • Not necessarily connected. • Dense for each connected components. • Connected overlay including links from a single node s • For every node x, preserves local reverse shortest paths to x •  • Preserves network-wide direct shortest path from s

  12. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • Path MPR Agenda • Motivation • Topology Pruning in Link State Routing • MPR for Topology Pruning • Path MPR • Impact of the Path MPR Modification

  13. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Path MPR (1) Selection of links to advertise • Specified in RFC 5449 Path MPR Selection Cost-Coverage Translation MPR Selection N(x) N2(x) N’(x) N2’(x) PathMPR(x) • N’(x) = {1-hop neighbors of x for which the direct link from x is locally optimal} • N2’(x) = {1-hop and 2-hop neighbors of x for which the local shortest path to x • has 2 hops}

  14. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • Connection: • Preservation of local reverse shortest paths: Path MPR (2) Resulting Overlay • Specified in RFC 5449 • For a node s of the network, the Path MPR overlay consists of: • N(s) • x in the network, links from x to PathMPR(x)

  15. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 2 1 1 4 1 3 1 3 3 1 1 5 6 Path MPR (3) (Non) Preservation of shortest paths • Example of non-preservation of shortest paths with general costs N(1) N2(1) N’(1) N(1) = {2, 3, 6} N2(1) = {4, 5} N’2(1) N’(1) = {2, 3} N’2(1) = {4, 5} PathMPR(1) = {3} Shortest_Path(41) = 4-2-1 • (with unit costs shortest paths are always preserved)

  16. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Path MPR Selection N(x) N2(x) E2x N’(x) N2’(x) E2x’ Cost-Coverage Translation MPR Selection PathMPR(x) Path MPR (4) Proposed correction • Idea: exclude links not participating in shortest paths (E2x)’ = {Links taking part in local reverse shortest paths to x}

  17. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 N’(1) 2 1 2 1 1 1 4 4 N’2(1) 1 1 3 1 3 3 1 3 3 1 1 3 1 1 5 6 5 6 Path MPR (and 5) Proposed correction • Application on the example N(1) = {2, 3, 6} N2(1) = {4, 5} N’(1) = {2, 3} N’2(1) = {4, 5} (E2x)’ = E2x \ {(6,1), (4,3)} PathMPR(1) = {2, 3}

  18. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 • Impact of the Path MPR Modification Agenda • Motivation • Topology Pruning in Link State Routing • MPR for Topology Pruning • Path MPR • Impact of the Path MPR Modification

  19. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 A first impact estimation (1) Distance-based costs + Proposed modification Path MPR (RFC 5449) 

  20. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 A first impact estimation (2) Random costs + Proposed modification Path MPR (RFC 5449) 

  21. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Conclusions & Future Work • General contributions • Conditions for a distributed topology pruning algorithm • Characterization of the MPR overlay: density, connection and preservation of shortest paths • Analysis of the Path MPR algorithm (RFC 5449) • Sub-optimality in non-unitary metrics • Proposed correction: proof of correctness and impact evaluation in static ideal network graphs • Future Work • Implementation and evaluation in more realistic ad hoc deployments, with different (non-hop) metrics

  22. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Questions ? E-mail: cordero@lix.polytechnique.fr

  23. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Backup Slides

  24. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Multi-Point Relays Heuristics Input: x, N(x), N2(x) MPR = {} MPR  {n  N(x) :  m  N2(x), m only covered by n} while ( uncovered 2-hop neighbors)MPR  n  N(x) : covers max. # of uncovered 2-hop neighbors Output: MPR(x, N(x), N2(x))

  25. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 1 2 3 4 (k-1) 2k-1 2k Examples of MPR overlay disconnection • Example of disconnected MPR overlay in a k-diameter network (k>0)

  26. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Path MPR Definitions

  27. MPR-based Pruning Techniques for Shortest Path Tree Computation IEEE SoftCOM 2010 Static Simulation Parameters • (Simulations run with Maple V) • Samples / experiment : 20 • Radio range (rr) : 150 m • Square grid length : 400 m • Node distribution : Uniform • Mobility : Static • Channel model : Ideal • Metrics model : K = 10

More Related