OSI layers 2 and 4 responsibilities:

LBC90805

Studying for the CCENT and reviewing the OSI have brought me some confusion again. Especially in the subjects of Error Recovery, Acknowledgment and Flow control. In the OSI model, doesn't the Transport layer take care of Error Recovery, Acknowledgment and Flow control? I have been making some flash cards and have seen referenced in several places the following concering LLC, the upper layer of the Data Link Layer. I quote from Wikipedia.org;

"The uppermost sublayer is Logical Link Control (LLC). This sublayer multiplexes protocols running atop the data link layer, and optionally provides flow control, acknowledgment, and error recovery."

For the CCENT as far as Error Recovery, Acknoledgment and Flow control are concerned, all of this should be taken care of by the Transport Layer?

Cessation

I used to be confused with the whole recovery part.
L4 does error recovery and flow control.
L2 just checks for the errors.

LBC90805

Well my understanding is Layer 2 does error checking and if there is an error drops the frame with the encapsulated data. Then Layer 4 doesn't receive Acknowledgment and tries to recover, not Layer 2.

Cucumber

The quote says it all, "...optionally provides flow control, acknowledgment, and error recovery". Notice the word optionally. The OSI layering model is flexible enough to tell you "this layer should do this and that, and maybe that too, and this other layer could do that as well". Its not written in stone, its just general guidelines.

Yes, Layer2 can optionally do Flow Control, ACK and ErrRecovery, Layer4 should do it, but even some Layer4 protocols dont do this at all (UDP), furthemore, some Application layer protocols do Flow Control (i.e.RTP), ACK and ErrRecovery as well.

Some old Layer2 protocols used to do all that stuff, i.e. some implementations of 802.2 used to do this, but its not common nowadays.

As time has passed by, "newer" layer 2 protocols have dropped these responsabilities. Why? Because network designers realized that Layer2 protocols should send frames as fast as they can rather than spending time doing chores that an upper layer could do. Layer2 should keep up with increasing speeds capabilities from Layer1. Thats why Frame Relay, current Ethernet implementations and ATM just do Error Checking, not even error recovery, just as LBC90805 said, they just drop bad frames and thats all.

So in summary, you have to learn what layers do and what they optionally can do. And bear in mind some layers can do the same activities other layers may do.

EdTheLad

Cucumber wrote:

Thats why Frame Relay, current Ethernet implementations and ATM just do Error Checking, not even error recovery, just as LBC90805 said, they just drop bad frames and thats all.

ATM can do single bit header error correction, i'd call this error recovery.

ITdude

Picky, picky.