Doughboy,
The terms "level" and "layer" are sometimes used interchangeably in communications circles, but when discussing the protocol stack defined by the International Organization for Standardization's (ISO's) Open Systems Interconnection (OSI) Reference Model (RM), the most commonly used (hence the proper) term is "layer." Put them all together, and these three letter acronyms are known collectively as the ISO/OSI-RM or simply the OSI-RM. Be patient, I'll get to your question...
L3 or Level 3 (LVLT) took some license in selecting their name, IMO, and their name refers to the Layer 3 of the OSI RM which is that suite of protocols that has to do with the Internet Protocol, or IP, which is LVLT's stated claim for being. TCP/IP is actually a combination of the OSI Layers 3 and 4 functionality, as shown in the chart below.
There are other companies such as Layer 0 and Level 1 (or is it Layer 1?) whose purposes surround the particular attributes of the protocols in those layers, as also illustrated in my makeshift chart below.
The OSI-RM actually consists of 7 Layers, usually read from the bottom up (where the physical layer feeds upwards until eventually you get to the actual application layer):
Layer 7...........Application
Layer 6...........Presentation
Layer 5...........Session
Layer 4...........Transport (TCP/UDP)
Layer 3...........Network (IP)
Layer 2...........Data Link (Frame Relay/ATM/Ethernet/Token Ring/FDDI...)
Layer 1...........Physical Medium (Copper/Fiber/free space)
See the slide at ieee-occs.org for a better graphic representation of this. In fact, if you have some time, the entire presentation at this site would be worth anyone's while if they are interested in TCP/IP standards development processes.
ieee-occs.org
The name of the presentation is:
TCP/IP A Tutorial On A Family of Protocols and Some Other Interesting Things: A Presentation for the IEEE OC Chapter Communications Society Joint Meeting with OC Chapter Computer Society January 27, 1997
=========
The other part of your question dealt with the edge.
The edge of any network is that portion where the subscriber base interfaces to the internal backbone, or core network components. In the public network, it could be considered the interface or "layer" (there we go again with layers) between the intercity transmission elements and the subscriber loops that feed residence or business. The latter would be considered the distribution network in many circles.
To rough it out then, if we followed a call from my home in Brooklyn to my brother John's home in CA,, it would take the following path
Franks' phone to 1st C.O. (distribution)
1st CO to tandem and Interexchange switches (edge]
NY Interexchange office to LA Interexchange office (core)
LA Interexchange to tandem switch to John's CO (edge)
John's CO to his home (distribution)
There is actually another couple of layers here, like the home distribution network or the personal service area PSA and the like, and the Full Service Area Network FSAN (a part of the future all optical distribution network) which is outside the realm of this discussion.
So, when 3Com says that they are catering mostly to the edge, they mean that they are serving that part of the network that has to do with dial up (and other means of direct and indirect access) connections to ISPs and other service providers.
More specifically, they are provisioning ISP POPS with dial access arrangements (dialup muxes, DSLAMS, concentrators, etc.) that are attached with T-1s to end office switches and other similar configurations, all of which takes place in what is said to be the network edge.
The core exists on the other side of the ISP POP, upwards towards the NAPs, or national access points. The NAPs are the main peering locations on the Internet. Collectively, the NAPs matrix of high capacity connections [T3, OC-12, OC-48, etc. whether ATM or IP )can be referred to as the Internet's core, so to speak, and this is where 3Com would most likely defer to CSCO and BAY, and a growing number of startups, soon, very soon.
HTH, Frank Coluccio
|
