Archived from groups: comp.dcom.lans.ethernet (
More info?)
thanks both
"Walter Roberson" <roberson@ibd.nrc-cnrc.gc.ca> wrote in message
news:d9k2im$efn$1@canopus.cc.umanitoba.ca...
> In article <TrWdnbT-38I7lCDfRVnygQ@scarlet.biz>,
> Chris <sdqfsd@qdvqsd.cv> wrote:
> :Assume that several packets with a total amount of 100 mbits (not in
bytes
> :to simplify) are send together . Assume that the theoretical speed of 100
> :Mbits/sec is a fact.
>
>
oes it mean that they are sent all together (in a kind of parallel
> :channels) and that the total broadband is busy for a second or are they
sent
>
ne after one but the sum of the busy time of all the packets will be
one
> :second?
>
> Neither -- the time required to send 100 Mbits of packets over
> a 100 Mbit/s link is more than 1 second, due to the mandatory
> IFG (intra-frame gap) of 96 bit-times. Even if you include the size of
> the preamble and CRC in your 100 Mbit count, during the IFG nothing
> is being sent. Including preamble and CRC in the bit counts, you
> can get slightly over 99 Mbits through in one seconds of 100 Mbit/s.
>
>
> Different signalling methods are used over different media, some
> involving discrete bits and some involving phase analysis of
> particular sample points of analog waveforms. In the analog case,
> considering the error correction methods and other fine details,
> bits are not transmitted independantly of each other but rather in
> small groups. (The interconnection is noticably stronger for gigabit.)
> Thus the question cannot be resolved without reference to a particular
> media -- and to a particular definition of what it means to
> transit bits in parallel.
>
>
> There are numerous good sources of information on the 'net. One
> with a useful (but not up-to-date) overview is
>
http://ckp.made-it.com/ieee8023.html
> --
> "I want to make sure [a user] can't get through ... an online
> experience without hitting a Microsoft ad"
> -- Steve Ballmer [Microsoft Chief Executive]