FCS in Hubs and switches

[Donnie80]

Just a quick question on clearing up something that I'm unsure about. Do Hubs and switches etc. carry out a FCS (i.e. crc32) on data or is it just left to the NIC's to calculate?

Regards,

D

 

[jimbopalmer]

i know that my Nortel Accelars and Baystack 350 & 450 switches validate FCS for every port.

i think most 'store and forward' switches do, and 'cut-through' switches don't

a store and forward switch gets the whole packet before transmiting any of it and a cuthrough switch starts sending before if finishes receiving, which is faster but more error prone.

I tried to remain child-like, all I acheived was childish.

 

[bcastner]

I think hubs do not, store and forward switches do.

 

[Donnie80]

OK, thanks guys, I thought that was the case.

Just a further point I'd like your opinion on:

Is there really any point in a switch carrying out a FCS? Wouldn't faults be detected by the NIC anyway?

Wouldn't the switch just pass on the fact that there is a bad FCS to the intended NIC along with the frame?

I'm new to this so please forgive if i'm asking stupid and/or irrelavant questions.

Regards

D

 

[manarth]

it would cut down on bandwidth if bad packets weren't retransmitted.

<marc> i wonder what will happen if i press this...[ul][li]please give feedback on what works / what doesn't[/li][li]need some help? how to get a better answer: faq581-3339[/li][/ul]

 

[jimbopalmer]

besides owering the number of bad packets relayed through the network, it is nice to have a record of which ports are getting the most errors


I tried to remain child-like, all I acheived was childish.

 

[Accessdabbler]

The switch may drop a bad packet but, to my knowledge, it doesn't inform the sending NIC that the packet was dropped. It would be up to the receiving NIC to tell the sender that something is missing.

 

[bcastner]

Accessdabbler is correct. The switch only cares whether the CRC is correct and it has an ARP entry to resolve the destination MAC address.

It is up to the protocol to have the sender retry.