This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


[Xen-devel] Re: [advices] patch for dynamically changing vif rate via pr

To: Jeremy Fitzhardinge <jeremy@xxxxxxxx>
Subject: [Xen-devel] Re: [advices] patch for dynamically changing vif rate via proc file
From: ahmed amamou <ahmed.amamou@xxxxxxx>
Date: Sat, 4 Sep 2010 01:22:52 +0200
Cc: "chrisw@xxxxxxxxxxxx" <chrisw@xxxxxxxxxxxx>, "Xen-devel@xxxxxxxxxxxxxxxxxxx" <Xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Tue, 07 Sep 2010 07:51:09 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <4C814FEA.5030604@xxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <AANLkTinph=1Ua7gSV845gUBOkfuUnmw43XX9R7jq04EE@xxxxxxxxxxxxxx> <4C814FEA.5030604@xxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx

Hi Jeremy thanks for your answer,

I�m working in network virtualization especially in link virtualization we are trying an approach in which a link is a combination of up/down link speed between two points.

But as we work in virtual environment fixing up/down link speed for each connection is not a good way o enjoy virtualization benefits. So we have as objective to implement an algorithm that decides each link speed according to priority accorded to each network.

We want to change VIF speed instead of doing network shaping because we follow a simple rule in networking, when you don�t need a packet throw it as soon as possible and because in this way we are protocol independent (major traffic shaping method are for IP one ).

For using proc is the only way I found for the moment to do what I want because I don�t really understand, at least for the moment, the way all this is working.

The best way to do it , I think is to force xen_netif to read credit and usec parameter from xenstore each time it verify tx_credit_exceeded but I didn�t find how to do it (as I say I don�t have a clear vision to all the code)

Thanks again for your answer don�t hesitate to emit any critics if you see any incoherence in this idea.

Best regards


Xen-devel mailing list
<Prev in Thread] Current Thread [Next in Thread>