Energy Optimal Routing Protocol for a Wireless Data Network Easwar Vivek Colloborator(s): Venkatesh Ramaiyan, Srikrishna Bhashyam Department of Electrical Engineering, Indian Institute of Technology, Madras. Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 1 / 12
Outline Problem motivation A linear data network A linear program formulation Impact of communication overheads Contention as overhead Conclusion and future work Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 2 / 12
Problem Motivation Cellular Architecture Data Forwarding Model A wireless network in a disaster scenario lots of users but few gateways Cellular architecture may not be appropriate! ad hoc models with little infrastructure needed nodes can cooperate in routing and forwarding packets Data needs to be forwarded to the gateway data can be forwarded either by users or by relay nodes Energy optimal routing strategies for such networks Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 3 / 12
A Linear Network - Problem Illustration Linear slotted wireless network with n nodes equally spaced data aggregation model! Homogeneous traffic at intermediate nodes data aggregated at the sink node Tradeoff b/w long hops and short hops multihopping drains nodes closer to the sink direct transmission drains nodes farther from the sink Objective: Routing protocols that minimize per node energy (OR) maximize network life-time Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 4 / 12
A Linear Program Formulation Let d η (η = 2) be the energy spent per transmission for a hop distance d Define energy of node i, E i as E 1 := x 1 + x 2 + x 4 + x 6 E 2 := x 2 + 4x 3 + x 4 + 4x 5 E 3 := x 4 + x 5 + 4x 6 + 9x 7 The optimization problem is E := min(max(e 1, E 2, E 3 )) such that x 1 = 1, x 2 + x 3 = 1 x 4 + x 5 + x 6 + x 7 = 1 x i 0 Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 5 / 12
A Linear Program Formulation The LP formulation is min E such that x 1 + x 2 + x 4 + x 6 E x 2 + 4x 3 + x 4 + 4x 5 E x 4 + x 5 + 4x 6 + 9x 7 E x 1 = 1, x 2 + x 3 = 1 x 4 + x 5 + x 6 + x 7 = 1 x 1, x 2,, x 7 0 We assume there are no overheads with communication! Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 6 / 12
A Linear Program Formulation A simple optimal routing strategy predominantly multihopping rest through direct transmission Energy consumption is almost n all data may be routed through nearest neighbors direct transmission may be avoided! A simple routing strategy (if there are no overheads) route data through multihopping! strategy nearly optimal for large n Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 7 / 12
Overheads in Communication d η is the energy spent per transmission for a hop distance d Additional energy is spent during packet communication overheads in communication protocol overheads in transceiver electronics With overheads, optimal stragegy also changes Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 8 / 12
Overheads in Communication - Illustration Let K t be the transmission overhead (for a packet) E 1 := (1 + K t )(x 1 + x 2 + x 4 + x 6 ) E 2 := (1 + K t )x 2 + (4 + K t )x 3 + (1 + K t )x 4 + (4 + K t )x 5 E 3 := (1 + K t )(x 4 + x 5 ) + (4 + K t )x 6 + (9 + K t )x 7 The LP formulation is such that min E E 1 E, E 2 E, E 3 E x 1 = 1, x 2 + x 3 = 1 x 4 + x 5 + x 6 + x 7 = 1 Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 9 / 12
Overheads in Communication - Illustration With overheads, optimal strategy changes As K t increases, more data is transmitted directly! No simple routing strategy! number of nodes, communication cost and overheads influence routes Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 10 / 12
Contention as Overhead In most ad hoc networks, a distributed MAC schedules nodes e.g., IEEE 802.11 DCF MAC Distributed MAC protocol causes collision of packets retransmissions are overheads in communications LP formulation does not model contention loss all transmissions are scheduled orthogonally Direct transmissions: all nodes compete with each other nodes farther away spend more energy due to retransmission Multihopping: only neighbor nodes compete nodes close to sink spend more energy in retransmission Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 11 / 12
Conclusion and Future Work Multihopping is approximately optimal if overheads are insignificant Characterization of overheads in communication very important in designing routing strategies With a distributed MAC, schedule aware routing protocols are needed Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 12 / 12
Conclusion and Future Work Multihopping is approximately optimal if overheads are insignificant Characterization of overheads in communication very important in designing routing strategies With a distributed MAC, schedule aware routing protocols are needed Future work design of simple routing strategies with a distributed MAC models for 2-D wireless networks insights on positioning gateways in a network area (to reduce energy consumption) Energy Optimal Routing Protocol for a Wireless IUATC-T3, Data Network 26 January 2011 12 / 12