BEACONLESS GEOCAST PROTOCOLS ARE INTERESTING, EVEN IN 1D J oachim - PowerPoint PPT Presentation

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate”

F AIR MEDIUM ACCESS At any point in time, every node has then same probability to be the next to “activate” This assumption abstracts from different underlying collision handling techniques

CENTER-DISTANCE VS CENTER-DISTANCE-P

CENTER-DISTANCE VS CENTER-DISTANCE-P

CENTER-DISTANCE VS CENTER-DISTANCE-P I

CENTER-DISTANCE VS CENTER-DISTANCE-P I J

CENTER-DISTANCE VS CENTER-DISTANCE-P V I J

CENTER-DISTANCE VS CENTER-DISTANCE-P V M I J

CENTER-DISTANCE VS CENTER-DISTANCE-P V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD&CD-P: V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD&CD-P: V CDist ( v ) M d I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD&CD-P: V CDist ( v ) M d I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD: V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD: V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD: V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD-P: V M I R J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD-P: V CDist ( m ) M I R CDist ( j ) J

CENTER-DISTANCE VS CENTER-DISTANCE-P CD-P: V CDist ( m ) M I R CDist ( j ) J

OUR GOAL

OUR GOAL Analyze and compare heuristics

OUR GOAL Analyze and compare heuristics Develop theoretical model

OUR GOAL Analyze and compare heuristics Develop theoretical model • Quality measure: success rate and RecMess

OUR GOAL Analyze and compare heuristics Develop theoretical model • Quality measure: success rate and RecMess • Discrete time setting: packets sent in rounds

OUR GOAL Analyze and compare heuristics Develop theoretical model • Quality measure: success rate and RecMess • Discrete time setting: packets sent in rounds • Conflict resolution: fair medium access

OUR GOAL Analyze and compare heuristics Develop theoretical model • Quality measure: success rate and RecMess • Discrete time setting: packets sent in rounds • Conflict resolution: fair medium access Problem. Validate beaconless geocast heuristics within our model, and analyze success rate and RecMess under various scenarios.

T ODAY

T ODAY 2 scenarios in 1D: • Unbounded reach • Bounded reach

T ODAY 2 scenarios in 1D: • Unbounded reach Messages are sent from left to right, everybody can “hear” everybody. • Bounded reach

T ODAY 2 scenarios in 1D: • Unbounded reach Messages are sent from left to right, everybody can “hear” everybody. • Bounded reach Messages are sent from left to right. Each node can only hear from its r predecessors.

1D UNBOUNDED REACH SCENARIO

1D UNBOUNDED REACH SCENARIO

1D UNBOUNDED REACH SCENARIO

F LOODING IN 1D UNBOUNDED REACH SCENARIO 6 6 6 6 6 6 6

F LOODING IN 1D UNBOUNDED REACH SCENARIO 6 6 6 6 6 6 6

F LOODING IN 1D UNBOUNDED REACH SCENARIO 7 7 7 6 7 7 7

F LOODING IN 1D UNBOUNDED REACH SCENARIO 7 7 7 6 7 7 7

F LOODING IN 1D UNBOUNDED REACH SCENARIO 8 7 8 7 8 8 8

F LOODING IN 1D UNBOUNDED REACH SCENARIO 8 7 8 7 8 8 8

F LOODING IN 1D UNBOUNDED REACH SCENARIO 9 8 9 8 8 9 9

F LOODING IN 1D UNBOUNDED REACH SCENARIO 9 8 9 8 8 9 9

F LOODING IN 1D UNBOUNDED REACH SCENARIO 10 8 10 9 9 10 10

F LOODING IN 1D UNBOUNDED REACH SCENARIO success rate 100% RecMess = nk n nodes, k messages 10 8 10 9 9 10 10

1D BOUNDED REACH SCENARIO

1D BOUNDED REACH SCENARIO r

1D BOUNDED REACH SCENARIO r

F LOODING IN 1D BOUNDED REACH SCENARIO 6 6

F LOODING IN 1D BOUNDED REACH SCENARIO 6 6

F LOODING IN 1D BOUNDED REACH SCENARIO 7 6 1 1

F LOODING IN 1D BOUNDED REACH SCENARIO 7 6 1 1

F LOODING IN 1D BOUNDED REACH SCENARIO 7 7 2 1

F LOODING IN 1D BOUNDED REACH SCENARIO 7 7 2 1

F LOODING IN 1D BOUNDED REACH SCENARIO 8 8 2 2 1

F LOODING IN 1D BOUNDED REACH SCENARIO 8 8 2 2 1

F LOODING IN 1D BOUNDED REACH SCENARIO 8 9 3 2 1

F LOODING IN 1D BOUNDED REACH SCENARIO success rate 100% RecMess = O ( rk ) n nodes, k messages, range r 8 9 3 2 1

RESULTS: RecMess Unbounded reach Bounded reach scenario scenario Lower bound Flooding M-heuristic T-heuristic CD CD-P Delay-based

RESULTS: RecMess Unbounded reach Bounded reach scenario scenario Ω( k ) Ω( k ) Lower bound Flooding M-heuristic T-heuristic CD CD-P Delay-based

RESULTS: RecMess Unbounded reach Bounded reach scenario scenario Ω( k ) Ω( k ) Lower bound O ( rk ) Flooding nk M-heuristic T-heuristic CD CD-P Delay-based

RESULTS: RecMess Unbounded reach Bounded reach scenario scenario Ω( k ) Ω( k ) Lower bound O ( rk ) Flooding nk min { Mk, 2 rk } Mk M-heuristic T-heuristic CD CD-P Delay-based