Date: Tue, 2 Nov 2004 22:38:35 -0800 From: Luigi Rizzo <rizzo@icir.org> To: Pawel Malachowski <pawmal-posting@freebsd.lublin.pl> Cc: freebsd-ipfw@freebsd.org Subject: Re: [PATCH] burstable dummynet pipe Message-ID: <20041102223835.B42486@xorpc.icir.org> In-Reply-To: <20041031203558.GA49557@shellma.zin.lublin.pl>; 09:35:58PM %2B0100 References: <20041031203558.GA49557@shellma.zin.lublin.pl>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Oct 31, 2004 at 09:35:58PM +0100, Pawel Malachowski wrote: > Hello, > > Attached (poor) patch (against 5.3-RC1) implements pipes that can be > overloaded (exceed their bw) for a (short) period of time. > It adds yet another parameter for pipe: `burst x', x in (K)Bytes. > Should also work with dynamic (clonable via mask) pipes. > > Configuring `x bytes burst' means we are allowing to transmit x bytes > with increased (by default: ~2x) speed. After that, traffic passes > through pipe as usual (bw-limited), however, burst credit is slowly > (by default: ~10x lower than bw) accumulated when pipe is idle (low to tell the truth, the 2x and 10x thing seems to me a rather kludgy and arbitrary way to implement 'bursts'. The most straightforward way to implement this feature is to let the pipe be "burst" bytes ahead of the programmed rate, which results in a rather trivial change to the current code. If you really want to burst out data and recover credt at a different rate, these extra rates should be configurable. so if someone thinks of committing this patch in the current form, please at least note in the commit log that i am opposed to that. cheers luigi
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20041102223835.B42486>