xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Wei Liu <Wei.Liu2@citrix.com>
To: Rohit Damkondwar <genius.rsd@gmail.com>
Cc: wei.liu2@citrix.com, "xen-devel@lists.xen.org" <xen-devel@lists.xen.org>
Subject: Re: dynamically set bandwidth limits of a virtual interface
Date: Fri, 28 Dec 2012 10:35:56 +0000	[thread overview]
Message-ID: <1356690956.2917.7.camel@iceland> (raw)
In-Reply-To: <CAHEEu857jjBLACY=MNKzqEkpGbjXw1cbfqy9AOyR5fVsuOmagg@mail.gmail.com>

On Fri, 2012-12-28 at 07:46 +0000, Rohit Damkondwar wrote:
> On Thu, Dec 27, 2012 at 6:03 PM, Wei Liu <Wei.Liu2@citrix.com> wrote:
>         On Thu, 2012-12-27 at 08:46 +0000, Rohit Damkondwar wrote:
>         > Hi all. I want to set bandwidth limits to a virtual
>         interface
>         > dynamically(without restarting virtual machine). I have been
>         browsing
>         > xen source code 4.1.3. I looked into libxen
>         folder(xen_vif.c) and
>         > hotplug(linux) folder. Earlier in xen 3.0 , xenvif struture
>         > (driver/net/xen-netback/
>         > interface.c + common.h) and tx_add_credit function could be
>         used to
>         > modify rate limits. I want to change bandwidth limits
>         dynamically of a
>         > virtual interface in xen 4.1.3. Where should I look for in
>         xen 4.1.3?
>         >
>         > Please help.
>         >
>         
>         
>         Xen vif has a parameter called 'rate', I don't know whether it
>         suits
>         you.
>         
> The rate parameter only restricts one way traffic(probably only
> outgoing). 
>  

Yes, you're right. So you need two way shaping.

> 
>         Also, you can have a look at external tool like tc(8). My
>         vague thought
>         is that Vif is just another interface in Dom0, tc(8) should be
>         able to
>         traffic-shape Vif. 
> 
> 
> Don't you think using external tool may decrease the eifficiency ?. If
> xen itself has capabailities ( provided by tc tool ), wouldn't it be
> more efficient ? 
> 

Do you see significant performance degradation when using tc(8) or any
other tools alike? If so, do report with figures, it can help us
improve.

> I have used this tool. It is good. It serves my purpose. But wudn't it
> be better to include the bandwidth limiting capabilities in xen
> itself? I am not sure about this. Currently I am just browsing through
> the source code. What do u think ?
> 

TBH I'm not sure about this either. Again, comparisons and analysis of
bottleneck would be helpful.

> 
> I have seen  function "set_qos_algorithm_type" and paramaters
> (qos/algorithm type,qos/algorithm params, qos/supported algorithms) in
> vif class. Would they be useful ? Are they available only for XEN
> Enterprise ?
> 

Do you see those in libxen source code? I don't think they are in use
now.


Wei.

  reply	other threads:[~2012-12-28 10:35 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-12-27  8:46 dynamically set bandwidth limits of a virtual interface Rohit Damkondwar
2012-12-27 12:33 ` Wei Liu
2012-12-27 12:38   ` Pasi Kärkkäinen
2012-12-31 14:14     ` Alex Bligh
2012-12-28  7:46   ` Rohit Damkondwar
2012-12-28 10:35     ` Wei Liu [this message]
2012-12-30 12:08       ` Rohit Damkondwar
2012-12-31 11:30         ` Wei Liu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1356690956.2917.7.camel@iceland \
    --to=wei.liu2@citrix.com \
    --cc=genius.rsd@gmail.com \
    --cc=xen-devel@lists.xen.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).