Talk:Fair queuing

From Wikipedia, the free encyclopedia

"Bus" network topology This article is part of WikiProject Computer networking, an attempt to build a comprehensive and detailed guide to Computer networking on Wikipedia. If you would like to participate, you can edit the article attached to this page, or visit the project page, where you can join the project and/or contribute to the discussion.
B This article has been rated as B-Class on the assessment scale.
Mid This article is on a subject of mid priority within Computer networking for inclusion in Wikipedia 1.0.

Contents

[edit] Algorithm

The algorithm description is difficult to understand and may even be wrong. Can someone try to clean this up by defining some variables, adding some equations or pseudo code. Help! --Kvng (talk) 23:44, 5 June 2008 (UTC)

I think the description in the article is confusing, and I invented fair queueing. The original idea was to have one queue for each source IP address, and service them in rotation. That's easy to understand, but doesn't take into account packet length, or allow for weighting some sources or classes of traffic more than others. So there are fancier schemes. Here's a badly written lecture on the subject, [1], and here's a better one.[2]. If the "virtual finishing time" concept is introduced, the explanation here should be at least as good as the one in the second reference, which has some good drawings. I'd suggest explaining the basic approach clearly, and referring readers to references for the more detailed explanation of "virtual finishing time", which is specialist detail for implementors. --John Nagle (talk) 02:28, 6 June 2008 (UTC)
Thank you. The second reference helped a lot. Virtual time is a round number in a a hypothetical bit-by-bit round robin implementation. Virtual times need not be updated when new packets arrive in to the queue(s). Clever. I'll see if I can find time a bit later to improve the algorithm description in the article. --Kvng (talk) 15:18, 6 June 2008 (UTC)
Right. "Virtual time" is a way to implement weighted bit-by-bit round robin. I think the name is confusing, but it was used in some early papers. (There was a tendency in some academic computer science circles in the 1980s to use the terms "virtual" and "abstract" wherever possible.)
Did you see the hydraulic analogy in the second reference? Weighted fair queuing works like a series of tubes going into a junction, with conservation of liquid flow, more flow from the bigger tubes, and the rule that you can't switch inputs in the middle of a packet. Cute. --John Nagle (talk) 16:42, 6 June 2008 (UTC)
The hydraulics analogy was a good start but didn't get me all the way there. Hydraulics don't map directly to network behavior (e.g. increased pressure increases flow in hydraulics but not in networks). --Kvng (talk) 19:13, 7 June 2008 (UTC)

[edit] Name change fiasco

The name of this page needs to be changed to Fair Queuing to reflect the correct spelling, but I can't figure out how to change a page name. —The preceding unsigned comment was added by 192.76.82.90 (talkcontribs) 22:54, 14 August 2006 (UTC).

It's done with a move tab the top of the page. Perhaps you need to create an account and log in to have that ability. I'll do the renaming, and drop the capital Q also. Thanks for catching what everyone else has missed! JonHarder 12:55, 15 August 2006 (UTC)
It turns out there is already an article with that name, which means an administrator will have to perform the operation. I'll make the request. JonHarder 13:02, 15 August 2006 (UTC)
Problem is partly solved. The name is now Fair Queuing, which is the most common spelling on the internet. It harmonizes with the "Weighted Fair Queuing" article. Unfortunatley I did not succeed in removing the capitalization. (I moved the redirecting page "Fair queing" to a temporary address, and then the "Fair queuing" article to "Fair Queuing". I was not allowed to move to "Fair queuing". Finally the redirecting page was moved to "Fair queueing", and a new redirecting page "Fair queuing" was created.) Mange01 21:54, 15 October 2006 (UTC)
The following discussion is an archived debate of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the debate was move; fair queueing should suffice. -- tariqabjotu 02:27, 21 August 2006 (UTC)

[edit] Requested move

Fair QueueingFair queuing – Correct spelling and capitalization. JonHarder 13:11, 15 August 2006 (UTC)

[edit] Survey

Add "* Support" or "* Oppose" followed by an optional one-sentence explanation, then sign your opinion with ~~~~

[edit] Discussion

Add any additional comments

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

[edit] A. Demers, S. Keshav, and S. Shenker didn't invent fair queuing.

I generally try not to blow my own horn on Wikipedia, but Demers, Keshav, and Shenker didn't invent fair queuing in 1989. I iinvented it in 1985. [3][4]. It's one of the classic Internet RFCs, and there's a published paper. I coined the term fair queueing, which first appears in that paper. That paper also introduced congestion collapse and first applied the tragedy of the commons problem to networking. But since it's my own work, I shouldn't edit the article. I'd appreciate it if someone else would do so. Thanks. --John Nagle 22:01, 15 October 2006 (UTC)

Done. It is a pleasure to meet you. Personally, I would not restrain myself from using WikiPedia for publishing terms that I have coined. Mange01 22:22, 15 October 2006 (UTC)