So could this 'header transformation' be the process of encapsulating & decapsulating (or "tunneling") an IP datagram?
So for example: A workstation attached to Router A wants to send a packet to a workstation attached to Router C and A & C are connected through a Router B.
Router A would encapsulate it and send it to Router B where it's decapsulated.
Router B would then compares the packet's destination network address found in the header with Router B's routing table and finds that it will have to send it through Router C.
Router B then re-encapsulates it and sends the packet off to Router C. (decreasing the TTL)
Router C decapsulates the packet to it original form, see's that the destination address points to one of its attached workstations and sends the datagram off to it.
Hmm...did I forget anything? Meh, it's just an example. = )
In theory, doesn't this form of 'Tunneling' represent a type of peer-to-peer network or communication? So to reword the question it would be: "Explain how routers route packets by giving an example of networks (peer-to-peer) using header encapsulations\decapsulati
Now wouldn't this be a correct way of interpreting the question?
(I'm assuming it's not that complicated)
Main Topics
Browse All Topics
by: SMackPosted on 2003-06-04 at 17:26:16ID: 8653285
Well, as far as layer three goes, the only transformation I'm aware of is the Time-to-Live (TTL), which should be decremented per router hop.
In a VLAN environment, a router is usually not the first device to see the packet, so tagging (VLANs), or trunking(ISL or Dot1Q) shouldn't be considered a "routine" thing.
In an MPLS environment, it's concievable that the packet could be assigned an MPLS-specific route for traffic management, but that's not "usual" at this point either.
In an HSRP environment, there's some "magic" done with addressing, but that's primarily at Layer Two - At layer three, each router participating has a base IP address that is represented by a virtual IP address ... again, not necessarily a "usual" thing ... especially at a basic router class.
There are some Layer two transformations: The MAC address of the ingress port is exchanged for the MAC address of the egress port (MAC addresses are only significant on the native segment).
Routers don't change protocols; i.e., you can't switch an IPX packet into an IP packet (well, you sorta can - using a "tunnel" ... like a GRE tunnel): That's not considered a basic router function. If it comes in as an IPX packet (and the router has code to handle an IPX packet) it will leave as an IPX packet (perhaps encapsulated, but it's still basically an IPX packet). Any protocols that the router doesn't know about are usually dropped. Some routers can be configured to bridge protocols that it can't route.
That's about it for me.....
Good Luck, hope this helps a little.
Scott