Is There a Clear Answer

Oh man! Ad-blocking software has been detected! :'(

This website is run by the community, for the community... and it needs advertisements in order to keep running. Blocking our ads means your killing our stats!
Please disable your ad-block, or become a premium member to hide all advertisements and this notice.

Hey guys I have just begone to study and could really use some clarification regarding (error correction and flow control in the Transport layer and error notification and flow control in the Data Link Layer ).

To me there seems to be an overlap, perhaps it is my authors wording, perhaps it is just me....

I have the understanding that Layer 2 is performing error notification to the sending host.

Regarding error correction in the Transport Layer, unsure here unless it is performing a check within the OSI layer itself.

Flow control in two layers.

 

Oh man! Ad-blocking software has been detected! :'(

This website is run by the community, for the community... and it needs advertisements in order to keep running. Blocking our ads means your killing our stats!
Please disable your ad-block, or become a premium member to hide all advertisements and this notice.

No There is not a clear answer, for example if you take three different certification books, authored by three different authors, you will have three different perspectives.

 

Conclusion Transport Layer 4 (sending host) & Transport Layer 4 (receiving host) are performing there own error checking, because it is a direct communication from L4 to L4.

DLC is performing error notification and not error correction.

 

Hi, sorry I missed this. Been MIA a while. My understanding is this:

At layer 2, the only error checking is the Frame Check Sequence (FCS) or CheckSum (CS). This is calculated by whatever device builds the frame - typically the sender - and it is placed inside the frame near the end.
The end device (the recipient) will read the bits coming in and then perform its own check calculation to generate its own Check Sum. If the end device's check sum matches the one in the frame, that means the data got through intact. Great - so pass it up to layer 3!
However if the check sums don't match, the data is corrupt. So the frame is dropped.

"But wait a minute," You may say. "Dropped!? That's it?" Perhaps you're wondering why... Why doesn't it tell someone, or ask for a resend? The simple reason is that it can't. At layer 2 there are no IP addresses - no global address system - just layer 2 local addresses. LAN addresses. Layer 2 can't see beyond this. Now, the original sender might not be on the local link. It could be anywhere in the world. At layer 2 there will simply be no way to get a frame back to them. You might imagine a switch could scan the packet, try to figure out a return address or something, but really it can't - the data's corrupt, so anything in it could be wrong. Just no. At layer 2 if the data is corrupt the only real option is to drop the packet. Forget it and move on, hoping that if it was important - some higher layer will be watching out for it.


However, layer 4 - the transport layer - has a higher level view. Layer 4 sees past the LAN and uses the (global) IP address system. The Layer 4 protocol TCP also brings the idea of sequencing into the mix. It can use these to create "sessions" or "connections" (which are basically just conversations between itself and another computer that it can track). This way, it can track sessions in progress. It can watch out for packets arriving out of sequence, or packets that don't arrive at all.
So when TCP on my computer has set up a conversation (session) with some distant computer at IP address 10.20.30.40, and I get packet 51, 52, 54 and 55 in the conversation, TCP will spot that packet 53 is missing. It knows the sender is at 10.20.30.40 - it knows that packet 53, specifically, was missing - so it can send a request back to the original source to get that packet sent again.

Hope that clarifies things a little.

 

Hey Zig - Thanks for the reply.

I totally got caught up or tunnel vission on the layer. It went right by me "du" DLC Layer 2 has no way to request; the only available action is to drop the corrupt frame.

Am I understanding this correctly that layer 2 "error notification" come is the form of a dropped frame and there is no actual "notification message sent up the stack".

Back to flow control - hum it seems to me if layer 4 is performing flow control it is doing this with respect to higher layers and there data stream and layer 2 is actually performing flow control with respect to the amount and frequency of frames being placed on the physical layer 1.

I think the authoring of certification material is highly competitive industry and the various authors try to distinguish their perspective from the other author's material, but at the end of the day an apple is still an apple no matter how it is described. I find my self referencing other material on the same topic for clarification. Zig you might want to keep an eye out on your perspective, it might just show up verbatim in the N-06 series of books.