How does fragment-free know frame isn't corrupt?

binarysoul

We know with fragment-free forwarding checks first 64-bytes to decide if a frame is corrupt or a collision has occurred. I'm curious without CRC check how does it arrive at that conclusion?

Is it the general switch intelligence and/or IOS that does that?

Nightflier101BL

If the switch receives a frame that is under 64 bytes in size, it's discarded. The frame isn't being checked for corruption, only the size is checked. If it's under 64 bytes in size (called runts) it figures it was damaged and drops it. The problem with this theory is that you can still receive frames with CRC errors because you can still receive frames 64 bytes in size but still be corrupted.

All the heavy lifting of making these switching decisions rest on the integrated ASIC chips on the motherboard. It's acts like a second CPU to handle all the decisions quickly.