PERFORMANCE EVALUATION OF MEDIUM ACCESS CONTROL FOR MULTIPLEBEAM

PERFORMANCE EVALUATION OF MEDIUM ACCESS CONTROL FOR MULTIPLE-BEAM ANTENNA NODES IN A WIRELESS LAN Dhananjay Lal Vivek Jain

OUTLINE n n n n Introduction Collision Probability per Beam Analytical Modeling for Basic CSMA and Slotted Aloha Performance Evaluation Expanded Receive Rule (ERR) 4 -Way-Handshake CSMA Summary Future Work

INTRODUCTION n n Wireless LANs with omni-directional antennas, are afflicted by low communication range, and often poor performance in noisy environments. Switched beam antennas offer enhanced throughput: n n Multiple beams may be used in parallel; more the number of beams the greater the scope of exploiting parallelism. The collision probability per beam falls as beam directionality increases.

COLLISION PROBABILITY PER BEAM n n Probability that i nodes lie within a beam: Pi Probability of collision in a timeslot: Pc(i) M: Total number of nodes in a beam; g: Transmission probability in a slot

COLLISION PROBABILITY PER BEAM (Cont. ) g = 0. 1 M = 1: 80 The former factor that provides performance gain with smart antennas is their ability to simultaneously receive multiple packets, or reciprocally, transmit multiple packets.

ANALYTICAL MODEL n n n Node has a total of ß switched beams, each beam being a sub-channel Receiver node (sink) has M neighbors (sources), each one having packets for the receiver node A sub-channel is a succession of regeneration cycles Each cycle has n An idle period, I n A busy period, B If U is the time spent in useful transmission, the throughput S in a subchannel is Total throughput, St is given by

ANALYTICAL MODEL (Cont. ) n n n Find the initial throughput per beam (sub-channel), S, based on number of neighbors This represents the fraction of time devoted to uncollided transmissions per beam Refine this estimate of uncollided transmissions to actual throughput per beam, Sβ, by factoring in the TDD requirement (beam synchronization) n n TDD: Keep the antenna array in reception mode for one packet duration at a time … can rx from multiple beams but cannot tx on any beam Find the total throughput, St, by multiplying with the number of beams Time Slots in CSMA: propagation delay time, a Time Slots in Slotted Aloha : packet Tx and propagation, 1 + a

ANALYTICAL MODEL FOR BASIC CSMA

ANALYTICAL MODEL FOR BASIC CSMA (Cont. ) TP = 1+a+f • Assuming nodes are distributed uniformly over area, each sub-channel (beam) has the same average number of nodes given by, • The total offered load on the entire channel (sub-channels), G=(M*g)/a

ANALYTICAL MODEL FOR SLOTTED ALOHA Throughput per beam falls as the number of beams is increased, the beam synchronization losses are heavy

PERFORMANCE EVALUATION: BASIC CSMA THROUGHPUT Basic CSMA throughput per beam, M = 40

BASIC CSMA THROUGHPUT, M = 40 1 beam 2 beam 3 beam 4 beam

BASIC CSMA THROUGHPUT, M = 60 1 beam 2 beam 3 beam 4 beam

SLOTTED ALOHA THROUGHPUT M = 40 M = 60

EXPANDED RECEIVE RULE (ERR) n n The long term total throughput, St, can attain a maximum value of unity regardless of the number of beams, B, or the offered load, G as: inf Throughput Improvement in CSMA: Expanded Receive Rule (ERR) n n Allow the node to continue reception beyond the packet reception time from the initial start of packet reception by an additional time td. TP. Amounts to relaxing the tight synchronization that was enforced for TDD between transmission and reception.

EXPANDED RECEIVE RULE (ERR) (Cont. ) 3 Beams, M=40, Basic CSMA 4 Beams, M=40, Basic CSMA

BEAMS UTILIZATION Number of Simultaneous Transmissions for Basic CSMA/ERR With increase in td, however, the law of diminishing returns to scale sets in for larger t d.

4 -WAY-HANDSHAKE CSMA THROUGHPUT 3 Beams, M = 40 4 Beams, M = 40

SUMMARY n n n Evaluated the performance of CSMA and Slotted Aloha gives better throughput than CSMA at lower loads; for heavier loads CSMA gives better performance Isolated the factor that drastically impedes the performance of CSMA, and proposed ERR, a scheme that provides improvement by pushing throughput above unity, and making it sensitive to the number of beams.

FUTURE WORK n This analysis ignores the hidden terminals. Although the effect of hidden terminals reduces with directional beams, the precise impact on throughput remains to be analytically modeled and is future work.

REFERENCES n n Dhananjay Lal, Vivek Jain, Qing-An Zeng, Dharma P. Agrawal, “Performance Evaluation of Medium Access Control for Multiple-Beam Antenna Nodes in a Wireless LAN, ” Tech. report CDMC-0703 -01, ECECS department, University of Cincinnati, submitted to IEEE Transactions on Parallel and Distributed Systems. J. H. Kim and J. K. Lee, “Performance of Carrier Sense Multiple Access with Collision Avoidance Protocols in Wireless LANs, ” Wireless Personal Communications, pp. 161– 183, November 1999.

Questions ? ? ? THANKS !!!