what layer of the IOS model is CRC erro

Find more posts tagged with

Comments

well it is the OSI model
segments and packets have a checksum field in their headers. And, frames have a fcs field in it's footer.
So, i guess they could be doing a checksum check at the transport, network, and data link layer.
But, I'm pretty sure what your thinking of is the crc done at the data link layer to see if the frame has been received properly.

yrwins

thanks.

james43026

Also, the CRC could be checked at layer 1, if you are using TCP offload, or TCP Checksum offload, at which point a built in integrated circuit of some sort on the NIC itself would be processing TCP / IP instead of the CPU / drivers.

OctalDump

james43026 wrote: »

Also, the CRC could be checked at layer 1, if you are using TCP offload, or TCP Checksum offload, at which point a built in integrated circuit of some sort on the NIC itself would be processing TCP / IP instead of the CPU / drivers.

TCP Offload is still layer 3+, though. Just because it is done in the NIC, doesn't change this.

However, the 802.11 standards have a "physical layer" which included the PLCP header, which has a CRC. It's not 100% clear to me if this is the same as the OSI physical layer, though.

But back to OP's question - it's almost certain that the answer they want is layer 2, and they are referring to the FCS at the end of ethernet frame.

Checksums (not CRC's though) are also in IPv4 headers (not IPv6) and TCP headers. You could invent your own protocol for layer 3 or 4, and use CRCs. And of course, anything running on layers 5-6 could use whatever it likes as well. There's nothing inherent in the OSI model to prevent CRC's being used at any layer.

james43026

OctalDump wrote: »

TCP Offload is still layer 3+, though. Just because it is done in the NIC, doesn't change this.

However, the 802.11 standards have a "physical layer" which included the PLCP header, which has a CRC. It's not 100% clear to me if this is the same as the OSI physical layer, though.

But back to OP's question - it's almost certain that the answer they want is layer 2, and they are referring to the FCS at the end of ethernet frame.

Checksums (not CRC's though) are also in IPv4 headers (not IPv6) and TCP headers. You could invent your own protocol for layer 3 or 4, and use CRCs. And of course, anything running on layers 5-6 could use whatever it likes as well. There's nothing inherent in the OSI model to prevent CRC's being used at any layer.

You know, for some reason I was thinking of TCP offload as a layer 1 concept since it was being done in hardware and not software, but you are correct. After thinking about it for a bit, and comparing it to taking layer 3 decisions from software to hardware, the fact that it's a layer 3 concept doesn't change.