CS332 Ch. 28

CS332 Ch. 28

Spring 2014Victor Norman

Access delay vs. Queuing Delay

Q: What is the difference between access delay and queuing delay?

A: I think the difference is that access delay refers to the case where the packet is at the front of the queue but even then cannot be sent because the shared medium is being used by some other machine. This could happen at a host, router, or switch.

Largest Delays?

Q: Which specific delay causes, percentage wise, the most delay? The book says queuing delays. Is that true because it seems like server delays could get really large as well?A: If Comer says it, it must be true… No, I think you are probably right. Comer does say, “Queuing delays account for most delays in the Internet.” (emphasis mine)

Bandwidth, speed, capacity, throughput! Arrrggggh!

Q: I found part of 28.5 quite confusing with the terminology. Could you explain what Comer is saying? A: There are 2 basic measures: throughput and delay. Delay is measured in units of time. Throughput is measured in bits per second, that you can transmit or receive on a medium or through a network.

Possible to decrease delay?

Q: In section 28.5 Comer states that it is possible to increase throughput but impossible to decrease delay. Why? Couldn't you just upgrade the transmission medium (i.e. go from CAT5 to CAT5e or 6, or even fiber)?A: Look at page 475 where Comer clarifies: adding more capacity will not decrease propagation delay.You can lower delay (theoretically) by buying a different service from your ISP – like MPLS.

Advertised Bandwidth

Q: When a ISP advertises an internet connection, what is the bandwidth metric that they are actually advertising to you, good, effective, or optimal throughput? A: What do you think?I think it is optimal throughput.

Measuring network performance

Q: Comer mentions problems in measuring network performance, but doesn't TCP use "adaptive retransmission" which relies on measuring network performance?A: Yes, TCP measures the RTT and adapts itself using it. But, it uses an algorithm that averages recent RTTs, still weighting the most recent values the highest, but not completely relying on just the most recent values. It adapts much more significantly to timeouts (supposed lost packets).

Minimizing Jitter

Q: How can service providers minimize jitter in a network?A: I don’t know if they can control jitter at all. Jitter is theoretically related to congestion, so if there is persistent congestion at some router, they might load balance traffic, light up “dark fiber”, make more pairings with other ISPs, etc…

VoIP and Jitter

Q: How does voice over IP minimize jitter, especially if packets are lost for a while? A: I don’t think there is really anything a network protocol can do to manage jitter…

Is jitter just a network problem?

Q: Does the measure of Jitter depend more on the network's status and throughput or does it have more to do with what requests are made for what data?A: The jitter Comer is talking about is just network jitter. Probably inconsistency in servicing requests at a server would cause similar problems to clients.

Minimizing delay/jitter

Q: How can we minimize all the sources of delay without making tradeoffs elsewhere? Or is it even possible due to the physical limitations of hardware?A: Invent totally new physical media and new protocols… or…Break the laws of physics, captain!

Delay-throughput product

Q: Why is it useful to estimate the number of bits in a network (what is the importance of the delay-throughput product)?A: The delay-throughput product gives you a measure of the number of bits in transit at any time. If you can pump TONS of data into the network before any comes out, then if some is lost, you may have to retransmit all of that data.

Q 1

Q: What is network latency?1. Changes in delay and duration of the changes2. time required to transfer data across a

network3. amount of data that can be transferred, per

unit time.A: 2.

Q2

Q: What do we call the amount of data that can be transferred per unit time? And what units do we use to measure it?A: capacity, in bits per second.

Q3

Q3: What is "access delay"? (Hint: you see it when using an Ethernet NIC.)

A: The time for a packet to get access to the network.

Q4

Q: Is throughput a measure of capacity or speed?

A: capacity

Q5

Q: What is the capacity of a 10base100 Ethernet line?

A: 100 Mbps

Q6

Q: If Calvin has two links to the Internet (through US Signal), each at 500 Mbps, what is the capacity of our link to the world?

A: Hard to say. It could be 500 Mbps or 250 Mbps, or 1000 Mbps depending on the connection type, contract, etc.

Q7

Q: How does throughput compare to goodput?1. throughput is the same as goodput2. throughput is always less than goodput3. throughput is sometimes less than goodput

and sometimes more than goodput4. throughput is always more than goodput5. All of the above.A: 2.

Q8

Q: If all packets in a stream have the same delay, what is the value of the jitter?A: 0

Q9

Q: Throughput and delay are theoretically independent. Explain why they are practically dependent on each other.A: When congestion occurs along a route (at an "intersection"), then packets going along that route will be delayed and throughput will go down.

Q10

Q: What is significance of the delay-throughput product?

A: It is a measure of how much data is in transit at any one time. It is significant because if there is a connection problem, then all the data that is in transit at one time that has to be retransmitted.