02-10-2007 06:43 AM - edited 03-05-2019 02:16 PM
Hello,
can I have some information on why ethernet packet should be at least 64 bytes?
Thank you.
02-10-2007 07:12 AM
If I remember correctly:
1. Ethernet uses CSMA/CD, it requires a certain amount of time for collision detection to be effective.
2. Time for a station to listen for collisions is given by the relation:
(frame size)/(transmission rate)
3. The transmission rate for ethernet is constant(10Mbps)the time is directly propotional to the frame size.
Therefore to get the required time for CSMA/CD to work, it is imperative that we have a minimum frame size of 64 Bytes.
Undersized frames must be padded to the required length.
02-10-2007 07:40 AM
The 64 byte spec comes from the original incarnation of Ethernet, which operated on fat coaxial cable (about the size of a garden hose).
There is a concept called "bit time" that describes how long it takes for the electrical signal representing one bit of data to travel; it's not instantaneous from one end of the cable to the other.
In the original Ethernet, it took 64bytes (512 bit times) of payload to guarantee that the signal could reach from one extreme end of the cable (at maximum length) to the other.
Making sure that signal could reach the entire cable permits Carrier Sense, Multiple access, Collision Detection (CSMA/CD)enabled hosts (that are transmitting) to realize that more than one station was transmitting on the same cable at the same time, and to back off for another attempt at some (sort of) random time later.
Good Luck
Scott
02-10-2007 09:03 AM
Scott,
I liked your answer beter than mine so I rated.
You do exagerate about the garden hose though or maybe I have a rather thick garden hose at the moment (3/4 inch) :-)
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide