Presentation is loading. Please wait.

Presentation is loading. Please wait.

SOLAR Joy Ghosh, Sumesh J. Philip, Chunming Qiao {joyghosh, sumeshjp, Sociological Orbit aware Location Approximation and Routing.

Similar presentations


Presentation on theme: "SOLAR Joy Ghosh, Sumesh J. Philip, Chunming Qiao {joyghosh, sumeshjp, Sociological Orbit aware Location Approximation and Routing."— Presentation transcript:

1 SOLAR Joy Ghosh, Sumesh J. Philip, Chunming Qiao {joyghosh, sumeshjp, qiao}@cse.buffalo.edu Sociological Orbit aware Location Approximation and Routing A Random Orbit Model and its Parameters Living KitchenPorch Conf. Room CafeCubicle HomeSchool Outdoors Home Town City 2 Friends City 3 Relatives MANET Level 3 Level 2 Level 1 Intermittently Connected Networks Sociological Orbits SOLAR Variations: Ongoing Research Non-probabilistic – Geographic forwarding to hubs o SOLAR Sequential – to all hubs in sequence o SOLAR Simulcast – to all hubs simultaneously o SOLAR Multicast – to a multicast tree of hubs Probabilistic – Intermittently connected networks o SOLAR-P – forward to hubs in probabilistic order o SOLAR-KSP – K-shortest paths; store & forward routing Key Concepts Every user periodically visits a list of places of social interests (i.e., hubs) Can utilize such mobility information for location approximation and routing Examples (at right): User 1 (green), User 2 (blue) and User 3 (red) attending a conference User 3 queries User 2 for the hub list of User 1 User 3 sends data to User 1 Advantage of Macro-level (hub-based) sociological orbital mobility profile does not require continuous location monitoring does not depend on exact movement in time or space acquaintance-based soft location management captures probabilistic routing in MANET & other networks (e.g., ICN) Query Optimization – Subset of Acquaintances to query Acquaintance A i has a Hub list H i = {h 1, h 2, …, h m } where h i is a hub H = {H 1, H 2, …, H n } is the set of hub lists covered by A 1, A 2, …, A n C = H 1 U H 2 U … U H n is the set of all hubs covered by A 1, A 2, …, A n Objective: find a minimum subset H’ of H such that: This is a minimum set cover problem – NP Complete Possible solutions: Greedy Set Cover, Primal-Dual Schema, etc. Minimizes the number of queries and optimizes the cache size GeneralParameters Simulation time1000sTerrain size1000m X 1000m No. of nodesVary, (Default = 100)Radio rangeVary, (Default = 200m) MAC protocolIEEE 802.11Mobility modelRandom Orbit SOLARParameters Total hubsVary, (Default = 15)Hub sizeVary, (Default = 200m) Hub stay time50s – 100sIHO Timeout250s – 500s Hub list size2 – Total hubsInter-hub speedVary, (Default = 10m/s – 30m/s) Intra-hub pause1sIntra-hub speed1m/s – 10m/s TrafficParameters CBR connections 200 Random (5 packets each) Data payload512 bytes per packet Conference Track 2 SOLAR Simulcast: Location Query and Routing Conference Track 1 Conference Track 3 Cafeteria Lounge Conference Track 2 Posters Registration Exhibits (b) Geographic forwarding of data to destination Conference Track 4 Conference Track 1 Conference Track 3 Cafeteria Lounge Posters Registration Exhibits (a) Geographic forwarding of location query to acquaintance Hub Centers Research Issues: Routing Objectives: Maximize data throughput (under energy and memory constraints) Minimize control overhead (number of location queries/updates) Minimize number of logical hops required for each location query Minimize number of acquaintances maintaining throughput Minimize the end-to-end delay (location query + data delivery) Routing Variable: Cache size (number of acquaintances) Logical hop threshold (acquaintance to acquaintance lookup) Hub list discovery probability (reliability of location approximation) Optimization problems: What is the minimum cache size required to achieve a desired discovery probability within a fixed number of search steps? Given a fixed cache size, what is the minimum number of search steps required to achieve desired reliability? What is the probability of Hub list discovery within a fixed number of search steps given a fixed cache size? Performance of SOLAR vs. conventional protocols SOLAR achieves high throughput, low control (signaling) overhead, and reasonable delay (even for destinations far away) Laboratory for Advanced Network Design, Evaluation and Research (LANDER)


Download ppt "SOLAR Joy Ghosh, Sumesh J. Philip, Chunming Qiao {joyghosh, sumeshjp, Sociological Orbit aware Location Approximation and Routing."

Similar presentations


Ads by Google