Topology Management Framework for Highly Dynamic MANETs Ricardo Sanchez, Joseph Evans {rsanchez,evans}@ittc.ku.edu

1. MOBILE BASE STATION MOBILE HOST BACKBONE WIRELESS LINK LAST-HOP WIRELESS LINK Topology Management Framework for Highly Dynamic MANETsRicardo Sanchez, Joseph Evans {rsanchez,evans}@ittc.ku.edu Mobile Ad Hoc Networks (MANETs) Recap Motivation • Self-organizing, self-configuring Wireless Networks • No fixed infrastructure, no centralized control • All nodes are mobile, all links are wireless • All/some nodes function as “mobile routers” (or relay nodes) • Nodes join/leave network arbitrarily • Applications: battlefield operations, disaster relief, etc. • Physical models: flat and hierarchical (shown below) • What happens in Highly Dynamic MANETs? • Intermittent and fast link activations and failures cause high rates of topology state changes • due to internal/external node conditions such as sudden mobility, power failure, harsh weather, high load, etc. • Most protocol approaches to end-to-end routing in MANETs • typically tackle one condition: mobility, energy, …. • assume rates of changes of short (insignificant) duration • Embed complexity at the routing protocol level to operate end-to-end over ANY underlying topology – does not always work! • But routing performance is significantly affected by high rates of changes [as reported in the research literature] • => Can we improve the responsiveness/efficiency of MANET routing protocols under highly dynamic conditions? • Challenges • Dynamic topology • Mobility, terrain, weather, … • Limited resources • BW, energy, processing load, … • Problem: MANET routing over multi-hop paths is difficult! Research Ideas Topology Management Framework • Building blocks: resilient link/cluster level control protocols • Cross-layer interactions with Routing/MAC functional blocks • (i) Link-level control protocol • Operate over any two neighboring nodes using special beacons • Distinguish link quality/stability between neighboring nodes • Establish “safe” links between “stable” neighboring nodes based on vectorized node-level (relative) metric • Tune node’s transmission range according to “stable” condition • (ii) Cluster-level control protocol • Operate over selected “safe” links among specific “stable” nodes • Construct/maintain variable-diameter clusters using vectorized cluster-level (aggregated) metric • Rely on special structure for dissemination of control information to minimize impact of signaling overhead during merge/split ops • (i) + (ii) = foundation for versatile MANET routing protocol • Decouple network topology structure from routing problem • Provide better representation of topology to routing protocols • Distinguish between physical and logical representation • Organize logical topology into connected/non-partitioned clusters • Characterize dynamic nature of topology with vector metrics • Require positioning/velocity/signal-strength information plus more • Metric values are essential to measure link/node/cluster quality… • … but metric vectors provides also an estimation for stability • May offer better resource utilization and/or savings capabilities • Construct and maintain stable cluster-based topology • Use mechanisms to merge/split clusters with the aid of resilient control structure • Goal: development of resilient and adaptive cluster-based topology management framework