Lectures10-14

# Lectures10-14 - Paradox of Residual Life ECE-CSE 861:...

This preview shows pages 1–4. Sign up to view the full content.

1 ECE & CSE ECE-CSE 861: Introduction to Computer Communication Networks Ness B. Shroff ECE & CSE Lectures 10-14 ECE & CSE Paradox of Residual Life Formally: Let {N(t),t 0} be a Poisson Process with rate λ Let S 1 , S 2 ,… be successive arrival times, and S 0 0 The number of arrivals in [0,t] N(t) The time of the last arrival before t S N(t) The time until the next arrival after t = S N(t)+1 ECE & CSE Paradox of Residual Life To Prove: P{Y t u|N(s), s t} = 1-e - λ u , u 0 In other words what we want to prove is that the time until the next arrival is exponentially distributed with mean 1/ λ . Proof: For u 0: {Y t >u} = {S N(t)+1 -t > u} = {S N(t)+1 > u +t} = {N(t+u) – N(t) = 0} A t =age Y t residual lifetime S N(t) t S N(t)+1 ECE & CSE Paradox of Residual Life P{Y t >u | N(s), s t} = P{N(t+u) – N(t) = 0 | N(s), s t} = P{N(t+u)-N(t) = 0} (Ind. Increment) = P{N(u) =0} (Stationary Increment) = e - λ u , u 0 P {Y t u | N(s), s t} = 1 - e - λ u , u 0 Q.E.D. The Paradox of residual life is important in calculating the E(W) in the queueing system. Homework: Find the distribution of the age of the process. In other words find P{A t u} ? E[S N(t)+1 - S N(t) ] = ? as t E[S N(t)+1 - S N(t) ] 2/ λ as t

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
2 ECE & CSE Reminder: Queueing Notation Notation: Infinite Buffer: A/B/C old notation A/B/C/ Single server Multiserver queue Inter-arrival time Service time number of servers ECE & CSE Reminder: Queueing Notation Finite Buffer: A/B/C/D Infinite Buffer: A/B/C Example: M/M/1 (Arrival Poisson, Service exponential, 1- server) M/G/1 Memoryless General ECE & CSE Reminder: Queueing Notation (c) M/M/m (d) M/M/m/N (e) G/M/1 , etc… # of servers = m # of servers = m Buffer size = N ECE & CSE M/M/1 Queue Possion arrivals and exponential service time. If τ n is the service time of the n th packet (customer), then P{ τ n s} = 1 - e - μ s , s 0 Given that a packet has been in service for t seconds, what is the probability that it will wait in service for an additional s seconds. In other words find P { τ n > s+t | τ n > t}. μ λ Poisson
3 ECE & CSE M/M/1 system } { } { } { ) ( s P e e e t P t s P n s t t s n n > = = = > + > = + τ μ } { } , { } | { t P t t s P t t s P n n n n n > > + > = > + > For all s, t 0. ECE & CSE M/M/1 system Additional time a customer (packet) waits in service is independent of how long it has waited so far. Quantities of interest: o Queue length distribution: P n o Avg. Queue length = Σ n P n o Avg. time delay E(T) o Throughput γ P n is defined to be the steady state probability of having n customers in the system ECE & CSE M/M/1 system P n = lim t P n (t). P n (t+ Δ t) = f(P j (t)) Question: If at time t+ Δ t the queue is in state n, then what are the possible states the queue can be in at time t? Probability of n packets in system at time t ECE & CSE M/M/1 system Let n 1 a) At time t, Queue is in state n+2 or n-2 with probability o( Δ T) b) Queue is in state n at time t a) No arrivals or departure in (t, t+ Δ T] b) 1 arrival and a departure in (t, t+ Δ T] c) Any other scenario is o( Δ T) c)

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 03/18/2012 for the course ECE 861 taught by Professor Shroff during the Spring '11 term at Ohio State.

### Page1 / 13

Lectures10-14 - Paradox of Residual Life ECE-CSE 861:...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online