Fundamentals of Computer Networks ECE 478578 Lecture 9

  • Slides: 11
Download presentation
Fundamentals of Computer Networks ECE 478/578 Lecture #9: Collision Resolution and Collision Free Protocols

Fundamentals of Computer Networks ECE 478/578 Lecture #9: Collision Resolution and Collision Free Protocols Instructor: Loukas Lazos Dept of Electrical and Computer Engineering University of Arizona

Collision Resolution Protocols Assume collision between only two packets All new arrivals defer from

Collision Resolution Protocols Assume collision between only two packets All new arrivals defer from transmission Node transmits at the next slot with prob ½ Let X denote the # of slots req to resolve collision Throughput = 2/3 2

Tree Splitting Algorithm Set S Resolution of collision of three stations Slot Xmit Set

Tree Splitting Algorithm Set S Resolution of collision of three stations Slot Xmit Set Waiting Set 1 S - e 2 L R e 3 LL LR, R 1 4 LR R e 5 LRL LRR, R 0 6 LRR R e 7 LRRL LRRR, R 1 8 LRRR R 1 9 R - 0 collision idle Feedback L LL collision success LRL R LR collision idle LRRL success New arrivals accumulate during contention period Throughput approx 0. 43 p/s (need 43 slots for 10 packets) collision LRRR success 3

Improvements Set S idle Two collisions in a row, R must have a small

Improvements Set S idle Two collisions in a row, R must have a small number of packets Incorporate to new arrivals Guaranteed to have a collision here collision L LL collision success LRL R LR collision idle LRRL success collision LRRR success 4

First Come First Serve Splitting allocation Split right away two collisions allocation waiting current

First Come First Serve Splitting allocation Split right away two collisions allocation waiting current time k current time allocation k+1 current time allocation k+2 current time k+3 5

Multi-Access Reservations Idea: Use reservation (contention) slots to make reservations for later transmissions Reservations

Multi-Access Reservations Idea: Use reservation (contention) slots to make reservations for later transmissions Reservations much smaller size than data Any method can be used to contend for reservation slots (or not contend) 6

Example: Satellite Reservation System Contention lasts N slots (equal to # of stations) Station

Example: Satellite Reservation System Contention lasts N slots (equal to # of stations) Station “i” transmits one bit in slot i if it has a frame to transmit After all the N slots are completed, every station knows who has frames to transmit, start data transmission in that order Example: Assume there are 8 stations Lower numbered stations have to wait for more time Efficiency d/(d + N), d = length of data frame 7

Inefficiency of Reservation System What if the number of stations N is large? Nodes

Inefficiency of Reservation System What if the number of stations N is large? Nodes have to wait for a long time to send data Reservation period grows very large 8

Binary Countdown Protocol Each station is assigned a binary address of bits Stations broadcast

Binary Countdown Protocol Each station is assigned a binary address of bits Stations broadcast their address one bit at the time, if they have a frame to transmit Take the OR of all bits wanting to transmit (nodes know the outcome of OR based on electric property of medium) If node with higher address transmits, back off Drop off after slot 0 Drop off after slot 2 Winner Efficiency: 9

(Un)Fairness of Binary Countdown Stations with highest addresses always win Good for implementing a

(Un)Fairness of Binary Countdown Stations with highest addresses always win Good for implementing a priority scheme To make the protocol fair, addresses of hosts need to be dynamic 10

Implementing Fairness After node A transmits, all hosts with address less than A increase

Implementing Fairness After node A transmits, all hosts with address less than A increase their address by one Node’s A address is set to zero All nodes before A get a higher priority Round robin type of protocol 11