Jump to content

Talk:Deficit round robin

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Terrible explanation and english use. Please re-write. Give clear definition of deficit counter, explicitly state what "the number" refers to.

Complete re-write (15:47, 30 April 2015‎)

[edit]
  1. Shorter introduction, trying to be homogeneous with the ones of Fair Queuing, Generalized processor sharing, Weighted Fair Queuing
  2. Re-write details of the algorithms, and given some pseudo-code
  3. Add of a "Performances: fairness, complexity" section, where to give pro/cons of DRR wrt WFQ and WRR
  4. Add of references to Cisco and Juniper implementations
  5. Add of a "See also" part, trying to be coherent with the ones of Fair Queuing, Generalized processor sharing, Weighted Fair Queuing

MarcBoyerONERA (talk) 16:01, 30 April 2015 (UTC)[reply]

Thanks for the contribution. I have cleaned up this rewrite a bit. It was not clear that before cleanup the new version was an improvement so please recognise that a WP:REVERT would have been easier and valid way to fix this. The original article was not in very good shape so I thought improving your contribution was a better way to move things forward.
Please keep in mind that we do not have consensus on your assertion that fair queuing is a family of algorithms. I have given notice but have not had time to back that language out of your other contributions. Please do not introduce this into any new contributions until we have worked it out at Talk:Fair_queuing#Family_of_Algorithms.
Do you have a citation for the pseudo code you have introduced or is that something you wrote yourself? ~Kvng (talk) 23:37, 3 May 2015 (UTC)[reply]
  1. On cleanup/revert: Sorry, I do not understand what you mean about "WP:REVERT", but I agree that you have improved my contribution.
  2. On "family of algorithms": I have noticed that there is no consensus, so I have used the least common point. From what I understand, User:Nagle and myself both agree on the fact that FQ is the name of a family of algorithm. The difference is that I claim that it is also the name of an algorithm.
  3. On the pseudo-code, a reference can be done on the initial article "ref name="shreedhar", or the articles of Lenzini. I don't know where to put it.
  4. Last, you have removed the reference to fairness. ("As presented for fair queueing, there is no unique definition on what is "fair".) Could you give me the reason why?
MarcBoyerONERA (talk) 14:37, 4 May 2015 (UTC)[reply]

Use of temp variable "q" (Revision 06:56, 17 July 2015‎)

[edit]

I do not consider that the use of temporary variable q in place of queue[i] improve the readability ot the article. First, it is not coherent with the use of the DC[i]. Second, the semantic of the line 'q:= queue[i]' is unclear: reference or deep copy? Why would the instruction q.dequeue() modify queue[i] ? For these two reasons, I suggest to undo this change. MarcBoyerONERA (talk)

You're right, I have reverted this change. Ferdinand Pienaar (talk) 00:12, 21 July 2015 (UTC)[reply]

July 2015

[edit]

This page is a hoax. There is a radio station called DWRR at a FM station in the Philippines. — Preceding unsigned comment added by 180.191.150.123 (talk) 12:49, 7 July 2015 (UTC)[reply]

DWRR stands here for Deficit Weighted Round Robin (DWRR). The acronym DWRR is valid according to the content of this page. The point that DWRR refers

too to a radio is not a reason to propose the deletion of this article

Nicolaslabrot (talk) 20:23, 7 July 2015 (UTC)[reply]

Undid revision 937844737

[edit]

I have undid revision 937844737 since

  • I do not know any non-weighted version of DRR, so, DRR and DWRR are two names for the same thing. If there is any difference, please provide a reference.
  • DRR can also be used for packets of constant size. And adding "for variable packet size" does not ease the readability IMHO.

MarcBoyerONERA (talk) 16:22, 27 January 2020 (UTC)[reply]