Date: Wed, 3 Mar 2004 01:24:06 +0300 From: Gleb Smirnoff <glebius@cell.sick.ru> To: Randy Bush <randy@psg.com> Cc: freebsd-net@freebsd.org Subject: Re: My planned work on networking stack Message-ID: <20040302222406.GA29412@cell.sick.ru> In-Reply-To: <200403022216.i22MGVQE029400@cell.sick.ru> References: <20040302031625.GA4061@scylla.towardex.com> <20040302042957.GH3841@saboteur.dek.spc.org> <20040302082625.GE22985@cell.sick.ru> <20040302084321.GA21729@xor.obsecurity.org> <20040302085556.GA23734@cell.sick.ru> <20040302092825.GD884@saboteur.dek.spc.org> <20040302095134.GA24078@cell.sick.ru> <40449B8E.A48B39B0@freebsd.org> <20040302160902.GB26977@cell.sick.ru> <200403022216.i22MGVQE029400@cell.sick.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Mar 02, 2004 at 02:16:13PM -0800, Randy Bush wrote: R> > I do not insist that AS pathes in kernel are good idea. If you R> > show me an other way to get AS information when constructing R> > netflow exports in kernel, I'd be thankful. R> R> do we need to rediscover why flow export places a large processor R> burden on criscos, junipers, prockets, ...? Not because of AS path info, definitely. Netflow does route lookup anyway to get nexthop and route masks. If route lookup will return a pointer to a structure with one more field it will not introduce ant additional load. My experience shows, that most load in flow processing is caused by: 1) memory allocation, 2) expiry lookups. -- Totus tuus, Glebius. GLEBIUS-RIPN GLEB-RIPE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040302222406.GA29412>