Layer 2/Router Question

bbarrick

If the process of the router when deciding to route the frame or not is to check for errors, then check the destination data link layer address to determine whether or not the frame is intended for that router then what does a router do when it receives an HDLC frame?

I'm guessing the router has that port configured to process all frames as long as they pass FCS?

Are there many protocols that do not require the router to include a data link layer address when being encapsulated?

Souljacker

I can't really pinpoint what you are asking.

HDLC is only going to be on a point to point WAN link. Are you asking what happens when a router has to route an ethernet packet over that link?

boobobobobob

All these things you're talking about are layer 2 proccesses. The data link layer address, HDLC, FCS checking - all layer 2. I think you might be using router too liberally when you really mean switch. So the switch will check the FCS and if an error is detected the packet is dropped.

bbarrick

Basically yes, does a router just skip the data link layer address check when it receives that frame on HDLC or is it manually configured to do so? According to Odom a router knows that its HDLC so it doesn't send the address, so the process he talks about seems just a general process, or specifically for Ethernet frames.

I was just wondering if that was a manual config on most routers or is it programmed into its logic?

bbarrick

No, I'm reading about the routing process/logic of routers specifically.

Souljacker

bbarrick wrote: »

Basically yes, does a router just skip the data link layer address check when it receives that frame on HDLC or is it manually configured to do so? According to Odom a router knows that its HDLC so it doesn't send the address, so the process he talks about seems just a general process, or specifically for Ethernet frames.

I was just wondering if that was a manual config on most routers or is it programmed into its logic?

I'm still trying to understand what you are asking.

On Page 89 Odom mentions that HDLC does do an error check and drops the frame if errors occurred. That should answer what I believe is one part of your question. The other seems to have something to do with routing, and why would you think that a frame that arrives on an HDLC link would behave differently than any other link? The router gets the frame, looks at the intended IP address, sees it is connected to the LAN for it, sends out an ARP request and the packet goes on its merry way. The Switch logic takes it from there which is a different issue.

What page specifically are you having trouble understanding? That might help.

bbarrick

It's Chapter 16 of the 100-101 book. I'm understanding the general process of the routers logic just was wondering how it works on HDLC. Let me explain myself a little better.

On a typical Ethernet link a router receives the frame. Checks for errors. Then checks the l2 address to make sure that it is in fact intended for it to route to its next destination or discard it(ie broadcast frame). If those to things checks out it pr

bbarrick

It's Chapter 16 of the 100-101 book. I'm understanding the general process of the routers logic just was wondering how it works on HDLC. Let me explain myself a little better.

On a typical Ethernet link a router receives the frame. Checks for errors. Then checks the l2 address to make sure that it is in fact intended for it to route to its next destination or discard it(ie broadcast frame). If those to things checks out it processes the frame as usual.

On an HDLC link however there is no l2 address in the header because there is only one possible location to send that frame. So, when a router receives a frame on an HDLC link, is its process physically programmed to process anything that passes the FCS since there is no l2 address to check?

Souljacker

bbarrick wrote: »

It's Chapter 16 of the 100-101 book. I'm understanding the general process of the routers logic just was wondering how it works on HDLC. Let me explain myself a little better.

On a typical Ethernet link a router receives the frame. Checks for errors. Then checks the l2 address to make sure that it is in fact intended for it to route to its next destination or discard it(ie broadcast frame). If those to things checks out it processes the frame as usual.

On an HDLC link however there is no l2 address in the header because there is only one possible location to send that frame. So, when a router receives a frame on an HDLC link, is its process physically programmed to process anything that passes the FCS since there is no l2 address to check?

Routers check layer 3 addresses for routing decisions, not layer 2. Layer 2 is MAC addresses. I agree with the other poster - I think you are mixing concepts. A router will reply with its MAC address to an ethernet broadcast if it has a route outside of the network to the intended IP, but other than determining the next hop (on local LAN) it could care less about Layer 2.

bbarrick

I'm not talking about next hop I'm talking about the process to determine whether or not the frame is intended for the router in which it has arrived. It does check the header for the l2 address to make sure that it is intended for it to process. Then it strips it, checks the Table to send it on etc....

Souljacker

bbarrick wrote: »

I'm not talking about next hop I'm talking about the process to determine whether or not the frame is intended for the router in which it has arrived. It does check the header for the l2 address to make sure that it is intended for it to process. Then it strips it, checks the Table to send it on etc....

So you are asking how does a router know that a frame it received on a point to point link is meant for it? I wish I had that version of the book, because it's explained very clearly in the third edition. The very nature of it being a point to point link implies the router should process it.

As an aside, there have been a couple threads on here about the newest Odom book being confusing\misprint\etc. I wonder if that's what you are running into with the explanations you are reading.

bbarrick

Yes, sorry for the mass confusion.

I understand it adds the HDLC header which doesn't include the l2 address to check on the receiving end. I guess what I was trying to figure out was if this is done automatically by the router or in configuration? Are there other protocols that change the way in which the router checks an incoming frame before processing it in the same way the HDLC does?

I haven't gotten to that point I guess.

Souljacker

bbarrick wrote: »

Yes, sorry for the mass confusion.

I understand it adds the HDLC header which doesn't include the l2 address to check on the receiving end. I guess what I was trying to figure out was if this is done automatically by the router or in configuration? Are there other protocols that change the way in which the router checks an incoming frame before processing it in the same way the HDLC does?

I haven't gotten to that point I guess.

I actually feel a little bad about pushing Odom's stuff so hard. The third edition of the books are phenomenal and truly a triumph. I wonder if they rushed him to refresh for the update and that's how so many errors ended up in there? HDLC is a WAN type point to point encapsulation that runs over serial links. PPP and Frame Relay also are Point to Point WAN connections. They don't follow the same rules that Ethernet does.

bbarrick

I agree with you there. Although I really like the book and in the case of this thread I think my curiosity and eagerness is getting ahead of me.

If you look on Mr. Odoms Facebook page you will see an interesting post I made a couple of days ago about a pretty big error in this book. He said there was no error but I knew it had to be. It's in a pretty bad place as well, subnetting.

I've spent the last, little over a month just on the first half of his book. I've caught some of the errors before they were listed in the errata file. It is a shame that it does seem to have been rushed to print. I plan on finishing it and combing it as thoroughly as possible though, only after catching that subnetting error they weren't aware of, Ill be more cautious with the last half.

Souljacker

bbarrick wrote: »

I agree with you there. Although I really like the book and in the case of this thread I think my curiosity and eagerness is getting ahead of me.

If you look on Mr. Odoms Facebook page you will see an interesting post I made a couple of days ago about a pretty big error in this book. He said there was no error but I knew it had to be. It's in a pretty bad place as well, subnetting.

I've spent the last, little over a month just on the first half of his book. I've caught some of the errors before they were listed in the errata file. It is a shame that it does seem to have been rushed to print. I plan on finishing it and combing it as thoroughly as possible though, only after catching that subnetting error they weren't aware of, Ill be more cautious with the last half.

I actually did go there last night after this thread - I'm Josie Revisited with a question just below yours (about something unrelated). Once I saw your questions I realized you were the same person who caught that subnet mask error (though thankfully that was a simple typo. I'm more concerned about the lack of clear explanation on topics that really were explained pretty easily in the third edition).

Do you also have Lammle's book? He's probably updated as well and if something doesn't make sense in the Odom book, it might be explained better by Lammle.

bbarrick

Souljacker wrote: »

Do you also have Lammle's book? He's probably updated as well and if something doesn't make sense in the Odom book, it might be explained better by Lammle.

His 100-101 book out yet? I may pick it up, but I'm going to finish this one first. Was also thinking about getting Cisco's flash cards they just released.