From: Wade Edwards (wade.edwards@xxxxxxxxxxxxxxxxxxx)
Date: Fri Aug 23 2002 - 16:52:18 GMT-3
I do agree that it is not a simple answer to say Ethernet is synchronous
or asynchronous. There are many components of Ethernet that are
synchronous and others that are asynchronous, as there are in other
networking mediums like ATM.
I looked on dictionary.com for an answer to synchronous and here is what
it said: pertaining to a transmission technique that requires a common
clock signal (a timing reference) between the communicating devices in
order to coordinate their transmissions. And asynchronous is:
pertaining to a transmission technique that does not require a common
clock between the communicating devices; timing signals are derived from
special characters in the data stream itself.
>From these definitions of synchronous and asynchronous I would say that
Ethernet is asynchronous because it does not require a common clock
between communicating devices (i.e. external clocking source) but the
timing signals are derived from special characters in the data stream
itself (i.e. the preamble). This is only talking about the clocking and
not the access methodology.
This is what I found about Manchester encoding: A method of transmitting
bits which enables the receiver to easily synchronize with the sender.
To me this means that Ethernet starts sending the preamble of a frame
and the receivers synchronize with the Manchester encoded data stream.
This gives them time to synchronize before actual data (i.e. network
traffic) is sent.
I guess it is like everything else. It depends how you define it and in
what context you are talking about it in.
L8r.
-----Original Message-----
From: Przemyslaw Karwasiecki [mailto:karwas@bellsouth.net]
Sent: Friday, August 23, 2002 2:34 PM
To: Wade Edwards
Cc: Asim Khan; Michael Snyder; Michael Spencer;
ccielab@groupstudy.com
Subject: RE: ethernet
I don't know if Ethernet is Synchronous or not. But:
Synchronization of clocks in receiver and transmitter is achieved
by manchester encoding, not frame preamble.
Frame preamble is used to delineate frames.
Difference between RS232 and Ethernet is, that after start bit,
clock of the receiver is free running in RS232, hence you cannot
reliable send more then couple of bits. (6,7,8). On the other hand,
in Ethernet, receiver needs to synchronize its clock to the signal
from the wire. It is possible, because of manchester encoding,
and PLL circuity, keeping receiver clock in synch.
BTW -- this system has horrible overhead. It doesn't utilize
available bandwidth of transmission channel in efficient way.
There are many techniques addressing this issue.
For instance FDDI was using some kind of encoding with only 25%
overhead, and some extra signalization incorporated into encoding.
As far as comparison between access methods to Ethernet media
and other medias, you are basically right. But does it implies
automatically that Ethernet is Asynchronous?
I guess, we are trying to attach synchronous/asynchronous label
to some broad concepts, like Ethernet or ATM, which are basically
a sets of different, simpler components. Those components possibly
can be qualified as either synchronous or asynchronous.
But I am not an expert in semantics. :-)
Take care,
Przemek
On Fri, 2002-08-23 at 13:38, Wade Edwards wrote:
> But with that logic, modems (specifically the connection from DTE to
> DCE) use synchronous communication because the start and stop bits
> synchronize each character of data. The receiver sees the start bit
and
> synchronizes to the transmitter clock. The clocking is transferred in
> the encoded data.
>
> I just don't think of it that way.
>
> By the nature of Ethernet, as you start adding more and more devices
to
> that medium total throughput goes down because each device is trying
to
> access the medium in an asynchronous manner. If there is never any
data
> to transmit on an Ethernet network (with keepalives turned off) there
is
> no clock.
>
> On an ATM network, if there is never any data to transmit there is
still
> clocking and order.
>
> To me this mean asynchronous for Ethernet and synchronous for ATM.
>
> L8r
>
> -----Original Message-----
> From: Asim Khan [mailto:asimmegawatt@yahoo.com]
> Sent: Friday, August 23, 2002 2:43 AM
> To: Przemyslaw Karwasiecki; Michael Snyder
> Cc: 'Michael Spencer'; ccielab@groupstudy.com
> Subject: RE: ethernet
>
> I think ethernet is synchronous. The reason is in any
> synchronous transmission, the receiver uses a clock
> which is synchronized to the transmitter clock. The
> clock may be transferred by either:
>
> 1)A seperate interface circuite.
> 2)Encoded in the data (like Manchester encoding,AMI
> encoding).
>
> Now in ethernet an encoded clock is used.
>
> Regards.
>
> Asim Khan
>
> --- Przemyslaw Karwasiecki <karwas@bellsouth.net>
> wrote:
> > Ethernet is using something called Manchester
> > encoding.
> > It basically means, that in order to provide clock
> > synchronization
> > between frame transmitter and receiver, each zero is
> > represented
> > by sequence of 01 and each one is represented by 10.
> > By doing so, it makes it possible to maintain clock
> > synchronization
> > even in case frame contains long sequence of zeroes
> > or ones.
> > And, yes, before each frame, there is a short
> > sequence send
> > called preamble (but I believe this is layer 1 not
> > 2),
> > which makes it possible to delineate beginning of
> > the frame.
> >
> > Is it synchronous -- IMHO yes, but it depends on
> > definition
> > of the term synchronous.
> >
> > Przemek
> >
> >
> > On Thu, 2002-08-22 at 17:17, Michael Snyder wrote:
> > > Where's the clock?
> > >
> > > Believe every Ethernet transmission starts with a
> > series of one's and
> > > zero's sent before the packet header.
> > >
> > > This layer two header provides the clock. So it
> > it's async before the
> > > packet is transmited, and synced as the packet is
> > transmited.
> > >
> > > Does this help?
> > >
> > > I have a better question for you, is ATM sync or
> > async. Really? You
> > > don't think there's a sync'ed clock signal on the
> > fiber cables. About
> > > as clear as mud huh?
> > >
> > > -----Original Message-----
> > > From: nobody@groupstudy.com
> > [mailto:nobody@groupstudy.com] On Behalf Of
> > > Michael Spencer
> > > Sent: Thursday, August 22, 2002 1:04 PM
> > > To: ccielab@groupstudy.com
> > > Subject: ethernet
> > >
> > > Is ethernet synchronous or asynchronous?
> > >
> > >
> > > ---------------------------------
> > > Do You Yahoo!?
> > > HotJobs, a Yahoo! service - Search Thousands of
> > New Jobs
> > >
> >
This archive was generated by hypermail 2.1.4 : Sat Sep 07 2002 - 19:48:35 GMT-3