1 / 13

Void Traversal for Guaranteed Delivery in Geometric Routing

d 2. s. V 1. V 3. d 1. V 2. Void Traversal for Guaranteed Delivery in Geometric Routing. Mikhail Nesterenko Adnan Vora Kent State University MASS November 09, 2005. d 2. ?. Geometric Routing: Routing without Overhead. no tables : each node knows only neighbors

ikia
Download Presentation

Void Traversal for Guaranteed Delivery in Geometric Routing

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. d2 s V1 V3 d1 V2 Void Traversal for Guaranteed Delivery in Geometric Routing Mikhail Nesterenko Adnan Vora Kent State University MASSNovember 09, 2005

  2. d2 ? Geometric Routing: Routing without Overhead no tables: each node knows only neighbors no message overhead: message of constant size no flooding: only one message at a time per packet no memory: no info is kept at node after message is routed no global knowledge • static nodes • each node knows its global coordinates • sender knows coords of receiver • simplest approach: greedy routing • message carries coords of dest. • each node forwards to neighbor closer todestination • problem: local minimum • what if no closer neighbor? s d1

  3. F4 Face Routing [BMSU’99] face – continuous area of planar graph not intersected by edges observation – finite number of faces intersect source-destination line idea - traverse each face intersecting sd-line, switch to next face when encountered • to traverse a face select to be outgoing the next edge after incoming counter-clockwise optimization (GFG/GPSR) – use greedy, switch to face to leave local minimum, switch back to greedy after approach destination closer than the local minimum, proceed iteratively to use GFG need planar graph • unit-disk graph – each vertex pair is connected if distance is less than fixed unit assume – approximates radio model • can locally construct Gabriel or Relative Local Neighborhood planar subgraph -- guaranteed connectivity -- no extra communication required HOWEVER d2 s F1 F3 d1 F2

  4. Radio Networks are Not Unit-Disk [David Culler, UCB] • non-isotropic • large variation in affinity • asymmetric links • long, stable high quality links • short bad ones THUS

  5. What to Do with Non-Planar Graphs? • planarization removes edges useful for routing • irregular signal propagation forces conservative estimates of edge length • increases route size • requires greater node deployment density void – continuous area in (not necessarily planar) graph not intersected by edges if unit-disk based planar graphs are inadequate is it possible to apply the ideaof traversal to voids innon-planar graphs? s V1 V3 d1 V2

  6. Outline • memory requirement for traversal – intersection semi-closure • traversal of voids of non-planar graphs • simulation setup, examples, results

  7. Intersection Semi-Closure to traverse voids nodes need to have more information about surrounding topology Definition:neighbor relation N over graph G is d-incident edge intersection semi-closed if for every two intersecting edges (u,v) and (w,x) either • (w,x) N(u) and there exist path(u,w)N(u) and path(u,x)N(u) neither one is more than d hops; or • (w,x) N(v) and there exist path(v,w)N(v) and path(v,x)N(v) neither one is more than d hops Lemma: in a unit-disk a neighborhood relation is 2-intersectionsemi-closed if for every node u and everyedge (w,x) such that |u,w| < 1 and |u,x| 2/3it follows that (u,w)N(u) • modest requirements on surrounding topology ensure intersection semi-closure x path(u,x)<d 1 u v w

  8. s V1 V3 d1 V2 VOID Traversal Algorithm edge_selection follows segment of the edges that borders the void two parts • edge_changemessage sent to node adjacent to next segment edge, node selects beginning of next segment (next intersecting edge)the selection minimizes the currentedge segment • sends edge_selectionmessage to the other adjacent node to confirm selection and forward message to node adjacent tonext segment edge GVG – void traversal joinedwith greedy routing similarto GFG a c b g e f h d traversaldirection edge_change edge_change i void k j

  9. Simulation Setup and Memory Usage • implemented FACE and VOID traversal in Java and Matlab • uniform distribution random graphs • fixed area of 22 units • 50, 100, 200 nodes • connectivity unit 0.3, 0.25, 0.2 respectively • fading factors of 1, 2 and 3 • generated graphs and computed unit-disk subgraphs • only 1 out of 350 generated had a connected subgraph for factors 2 and 3 • generated connected unit-disk graphs and added extra edges according to fading factor memory usage • FACE – proportional to average node degreed • VOID – proportional tod f 1 probability f=1 f=2 f=3 u 2u 3u distance

  10. FACE vs. VOID: Example Routes • 50-node graph, fade factor is 2 • FACE: 13 hops • VOID: 11 hops

  11. VOID vs. FACE: Average Route Length • randomly generated 10 pairs of nodes for each graph • used paired comparison to estimate route length improvement • comparison based on (HopCountFACE- HopCountVOID)/HopCountFACE

  12. Future Work for degenerate graphs, to establish the neighborhood, the node has to explore sizable portion of the network • what are the practical criteria for limiting graph exploration? • how certain are we that all intersecting edges are discovered? • what are the adverse effects of missed edges on VOID? x v u w

  13. Void Traversal for Guaranteed Delivery in Geometric Routing Mikhail Nesterenko Adnan Vora thank you

More Related