At 5:57 PM +0000 3/14/03, Scott Roberts wrote: > > In the end, the device either routes or bridges the frames it >> receives, but takes no action that can be distinctly described as layer >> three switching. >> >> Pete >> > >to my basic understanding ALL routing has a switching component to it >already, whether we're talking about regular routers or L3 switches.
Correct for the Cisco product line. We do have some cases, especially in the service providers, where we have things like Route Servers. These speak BGP and manipulate and readvertise routing information, but have zero forwarding capability. These were once important in scaling eBGP at exchange points, but, as router processors have become more powerful, route servers are used more for traffic monitoring and research functions. RsD is one publically available implementation of a route server; it is a GateD derivative that runs on assorted *NIX platforms. There are also routing-only (i.e., forwarding free) devices used as protocol test and benchmark instruments. See http://www.ietf.org/internet-drafts/draft-ietf-bmwg-conterm-04.txt for a BGP-specific discussion. >process >switching, fast switching, autonomous switching, distributed switching, >etc... are all the ways the packets are moved between interfaces on a >router. therefore both layer 2 and layer 3 'switch' irregardless of the name >on the chassis. > > I personally view the sole distinction between the standard routers/bridges >and the multilayer switches as the use of ASICs. > People may gain a lot more insight if they go beyond the initial, marketingspeak of "oh, it's an ASIC" and actually understood what is _in_ an ASIC. Most routing ASICs have processor, memories, etc., just as does the main processor. They do tend to have specialized instruction sets, minimum memory access time because their RAM is on-chip or on-chipset, don't spend time doing instruction decode (unless they already are microcode sequencers), etc. A sufficiently fast RISC processing system may be as fast or faster than an early routing ASIC. The local memory in an ASIC can be a double-edged sword when you are doing things like multicasting, and no memory is shared among forwarding processors. Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=65466&t=65215 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
