On Mar 23, 2007, at 6:31 AM, Templin, Fred L wrote:
Clearly there is also baggage with the terms L3a/b/c, and L2.5 does not capture it because there is not any one clearly defined layer that occurs between L3 and L2.
they didn't describe that because of the concept of sub-layering, which was summarily tossed on its ear when IEEE 802 decided that it could work on what we now call 'ethernet switching' when all technologies that form networks were supposed to move to ANSI (early 1980's) if it worried about what header was used when forming the network.
L2 in the OSI model takes the signal delivered by the physical layer (the wire, fiber, or other channel, and the means by which data is encoded there) and interprets it into messages at that layer, permitting two devices that are physically connected to exchange information. L3 in the OSI model allows such L2 connections to be relayed to form networks. So from that perspective (duh) an ethernet- switched LAN is a network, and Ethernet is being used as an L3-of-sorts.
from my perspective, it is more useful to consider the network layer sub-layered into an intranet layer and an internet layer. X.25, ATM, Frame Relay, Ethernet Switching, MPLS, and so on have the characteristic that the use the lowest communication layer's header to form a network (what I call intranetworking), and IP transcends that to concatenate intra-networks to form the Internet. That is of course simplistic in the sense that ip-in-ip is used in various ways, so IP can also be an intranet of sorts. But it at least gives the bits rational descriptions.
The layer you're looking for is a sublayer of the network layer. _______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
