I am trying to get my grips around bandwidth and latency. I have read Stuart Cheshire's essay, however, there are a couple of paragraphs that I cannot understand.

Paraphrasing: Assume we are to compare two data link connections, 64Kbps ISDN and 1/150 share of a 10Mbps Ethernet. Then, say we have a game server that is transmitting 40K of data every 10 seconds or 32kb/sec. He states that this is only 50% capacity of the ISDN line.He goes on to say "150 users doing this on an Ethernet is only half the capacity of the Ethernet". How did he get that? That is, 150 users is 50% capacity of the Ethernet?

He then says that on the ethernet, the player receives the data in about 32 ms (320kb / 10Mb/sec). "If the server is not the only machine sending packets on the ethernet, then there could be a contention for the shared medium, but even in that case the average delay before the player receives the data is only 64ms" How did he get that? the 64ms?

Also, wondering if someone can explain the 2nd paragraph of page 8/11 (link below)

The link to the article is here (pg 8/11): http://rescomp.stanford.edu/~cheshire/rants/Latency.html

  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin Aug 8 '17 at 20:56

The logic seems to be if there is not contention I will get the data at line rate (32ms) 150 users woudl consume 150*40000*8 = 49,152,000 which is 50 % of 10 meg. If the link is contended and hence policed to 50 meg it will take twice as long as only 50% of the BW is usable

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.