Date: Sun, 17 Oct 1999 12:07:40 -0400 (EDT) From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> To: jonathan michaels <jon@caamora.com.au> Cc: freebsd-net@FreeBSD.ORG Subject: ethernet switch type and freebsd (Re: can two fast ethernet cards work in a freebsd box ?) Message-ID: <199910171607.MAA46454@khavrinen.lcs.mit.edu> In-Reply-To: <19991017163640.B24189@caamora.com.au> References: <199910161104.TAA26753@eembox.ee.ncku.edu.tw> <Pine.BSF.4.10.9910161424540.81531-100000@mail.wolves.k12.mo.us> <19991017090323.A23931@caamora.com.au> <38094456.B210ECEC@softweyr.com> <19991017163640.B24189@caamora.com.au>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Sun, 17 Oct 1999 16:36:40 +1000, jonathan michaels <jon@caamora.com.au> said: > i've heard a reasonable amount of chatter, when its been > raised, about switch "layer N". This refers to layers in the new-irrelevant-but-still-popular OSI Reference Model. Layer 1 is physical, layer 2 is link, layer 3 is network, and so on up to 7 (application). In the Internet reference model, the layers are subnetwork, internet, transport, and application -- this cuts out the useless OSI session and presentation layers. (Shows what happens when your network is designed by smart people who actually build things as opposed to a committee.) Historically, a packet switch operating at the subnetwork (OSI link) layer was called a ``bridge'', and one operating at the internet (OSI network) layer was called a ``gateway'' or more recently ``router''. About eight years ago, the term ``switch'' was revived, to refer to bridge architectures in which packet forwarding was implemented in hardware, analogous to ATM's cell switches, and more specifically such designs where the internal interconnection fabric had enough capacity to handle all bridge ports operating at full line rate simultaneously. More recently, numerous vendors have developed hardware router architectures with analogous behavior. It thus became natural to describe these routers as ``layer-3 switches'', as compared to bridging ``layer-2 switches'' -- and in fact most of the former also implement the latter. Finally, a whole new evilness has been visited on unsuspecting humanity by the appearance of so-called ``layer-4 switches''. Depending on which vendor you talk to, this may mean different things -- the classic example of what Debbie Deutsch of Lucent calls a ``marketechture''. To some vendors, a ``layer-4 switch'' is simply a layer-3 switch which can look at transport headers to provide Quality of Service functionality or packet filtering. To other vendors, a ``layer-4 switch'' is a specific kind of NAT kluge used for load-balancing across Web servers. All of these names principally serve the purpose of the vendors' marketing departments. I prefer to use terms like ``bridge'' and ``router'' if I am referring to a specific type of packet switch, and reserve the term ``switch'' for the generic. But then again, I'm lucky enough to work in a place where I can call something a ``gateway'' and still be understood.... > specifically, is thier support in freebsd fro different switch > 'layers' Yes and no. FreeBSD can act as a router (layer 3) or as a bridge (layer 2), but the machines on which FreeBSD runs on typically do not have hardware forwarding support or a sufficiently beefy bus architecture to back it up. On the other hand, there are several vendors who will sell you a switch which happens to run FreeBSD on its management processor, like Juniper. -GAWollman -- Garrett A. Wollman | O Siem / We are all family / O Siem / We're all the same wollman@lcs.mit.edu | O Siem / The fires of freedom Opinions not those of| Dance in the burning flame MIT, LCS, CRS, or NSA| - Susan Aglukark and Chad Irschick To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199910171607.MAA46454>