Date: Thu, 26 Aug 1999 12:01:46 -0700 (PDT) From: Matthew Jacob <mjacob@feral.com> To: "Kenneth D. Merry" <ken@kdm.org> Cc: "Louis A. Mamakos" <louie@TransSys.COM>, freebsd-hardware@FreeBSD.ORG Subject: Re: what's the best working gigabit ether card... Message-ID: <Pine.BSF.4.05.9908261200230.996-100000@semuta.feral.com> In-Reply-To: <199908261857.MAA84266@panzer.kdm.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Louis A. Mamakos wrote... > > > > > > Another thing to keep in mind, if you're going to be connecting more than > > > two machines, is that the Alteon switches are the only ones that I've seen > > > that currently claim to do jumbo frames. They cost a bundle, but they're > > > more or less the only game in town. My guess is that will change > > > eventually. > > > > Packet Engines is also doing jumbo sized Gigabit ethernet in their > > switches. This was something that was a requirement at work (UUNET) > > since the backbone already carries 4470 bytes frames, and we didn't > > want to have to fragment going over gigabit ethernet plumbing. > > That's very good to know, thanks. I wonder why they don't mention it on > their web page? > > Another odd thing is that they claim there is a 3rd party FreeBSD driver > for their Gigabit ethernet PCI card, although I don't know of one. > (There's certainly not one in the tree.) I've been (possibly mis-) informed that Packet Engine's GigEther card has been dropped. In fact, that's why I've been contacted to look into alternates. Essential Networks has a GigE variant of their RoadRunner card (HIPPI)- porting Kevin Lahey's driver from NetBSD to FreeBSD for that card has been on my list for a while... > > > You might also check around for an Internet Draft recently published > > on how to encapsulate jumbo-sized frames. This can be problematic > > for some protocols that use SNAP encapsulation since the ethernet > > type field is used as a length; this normally isn't a problem since > > the lengths were smaller then the range of ethernet types assigned. With > > jumbo frames, this is no longer true. > > > > In our case, this came up in the context of encapsulating CLNS frames > > on the wire. This is used commonly on ISP backbones that run Integrated > > IS-IS routing protocol as their IGP. > > Hmm, thanks for the info. > > Ken > -- > Kenneth Merry > ken@kdm.org > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.05.9908261200230.996-100000>