Problems with synchronization and clock recovery cause errors like...

Hi.

"Problems with synchronization and clock recovery on Ethernet cause errors like..."

I've got a question in a test and can't find what that synchronization drifts can do. Packet loss? Something else, more specific?

Thanks in Advance, Chris.

Possibly I am thinking of something different than you are, but the IFG is not 64 bit times but rather 96 bit times. 48 bit times would be half of the IFG not 2/3.

(snip)

As they are asynchronous, a station can start sending at any time.

The 2/3 rule removes any advantage to a device with a slightly faster clock. As each has its own crystal oscillator, they will never be exactly the same.

-- glen

Rich Seifert napisa³(a):

Thanks. I know the test's old, but I didn't do it, I just got to pass it ;)

Another question is about the interframe gap in CSMA/CD Ethernet. Why would a station make a decision to start sending in exactly 2/3 of an IFG (64bit times)? The minimum is 47BT in this configuration so why

All I know is that 2/3 of IFG ist the time it still listens for a present carrier.

Chris.

Yep, IFG is 96 BT, all I know is to know what conditions are met when a station sends in 2/3 of the IFG (it's not written '64', I just calculated 2/3 of 96BT so I got 64BT)

Answer could be : "it's a Gigabit Ethernet and a hardware vendor used the smallest IFG possible - 64BT) to get highest data rates, but a) 1000Base harldy ever operates in half-duplex (that's what I've heard) and b) 99% of this test ist about 10Base and 100Base so my answer is probably wrong :/

Chris.