Date: Tue, 20 Sep 2005 17:27:14 +0200 From: Jeremie Le Hen <jeremie@le-hen.org> To: vladone <vladone@spaingsm.com> Cc: freebsd-ipfw@freebsd.org Subject: Re: dummynet patch Message-ID: <20050920152714.GF24643@obiwan.tataz.chchile.org> In-Reply-To: <477488950.20050920130453@spaingsm.com> References: <97663975.20050917141303@spaingsm.com> <20050919122154.GM51142@obiwan.tataz.chchile.org> <477488950.20050920130453@spaingsm.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, > Yes, thanks! But is a little redundant and confused to pass packets to > multiple pipe and queue. Isn't more elegant to put an option on queue > that limit maximum bandwitdth to that queue (like "bw" option for pipe)? > I dont know programming (not well), but i think that, can do the job, > if is put an supplementary condition, to verify if bandwidth > allocated for that queue is less or great than an "bw" parameter. > An queue declaration like: > ipfw queue 1 config weight 10 pipe 1 bw 128kbits/s > ipfw queue 2 config weight 20 pipe 1 bw 256kbits/s > is more clear and efficiently. > This mean that an queue receive bandwidth according with they weight > but no more that value indicated by "bw" parameter. > Someone with experience and that know code for dummynet, can make easy > (i think) an patch for that. Many folks are reluctant to add syntactic sugar on IPFW when it does not add any functionnality. I think I am too : if we add every shorthand that one can think of, ipfw would become a real mess. Furthermore, pipes and queues are two distinct objects which have different semantics. Pipes are used to emulate a physical network link, with two main properties : bandwidth and delay, while queues provide WF2Q+ policy (see ipfw(8) manpage). Regards, -- Jeremie Le Hen < jeremie at le-hen dot org >< ttz at chchile dot org >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050920152714.GF24643>