WN - 3.05

Question 1

Describe the poisson process

Answer 1

Stochastic process in which events occur continuosly and independently of one another

Question 2

Explain burstiness

Answer 2

A bursty source generates traffic in random clusters

Question 3

Is deterministic traffic bursty or not bursty?

Answer 3

bursty

Question 4

What is the poisson process in discrete time

Answer 4

Bernoulli process

Question 5

When is the burstiness of poisson and bernoulli processes removed

Answer 5

Removed for aggregated traffic sources

Question 6

What is self-similar traffic?

Answer 6

Maintains its burstiness at any time scale

Question 7

Why is burstiness important?

Answer 7

1. Peak traffic demands on bugger resources can lead to overflow and lost traffic
2. Peak demands may create QoS problems in a network
3. Need to characterise burstiness for traffic sources in a QoS environment

Question 8

What are the properties of the self-similar phenomena?

Answer 8

1. Have structure at arbitrarily small scales
2. A self-similar structure contains smaller replicas of itself at all scales

• For real phenomena, properties do not hold indefinitely; however, they hold over a large range of scales

Question 9

What is the Hurst parameter

Answer 9

A key measure of self-similarity

When H=0.5 -> indicates the absence of self-similarity

H closer to 1 indicates a higher degree of persistence of long-range dependance

Question 10

What is data traffic well-modelled as?

Answer 10

A self-similar process in many practical networking situations

1. Ethernet traffic
2. WWW traffic
3. TCP, FTP traffic
   VBR video

Straightforward queuing analysis using poisson traffic assumptions inadequate to model this type of traffic

Question 11

In regards to ethernet traffic, what happens when the load increase

Answer 11

The Hurst parameter increases as well

Question 12

What does aggregating several streams do?

Answer 12

It does not remove self-similarity

Question 13

Poisson is not a good model in this case. What is?

Answer 13

Superposition of many Pareto-like ON/OFF sources

Question 14

Elaborate on the performance implications

Answer 14

1. If actual data is more bursty than originally modeled, then the original models underestimate average delay and blocking
2. Self-similarity leads to higher delays and higher blocking probabilities
3. Therefore, self-similarity leads to a poor fit with traditional queuing theory results