Re: AW: Deconstructing OAM&P
Walter,
I also used to support a very large Token Ring environment. Source Route
Bridging worked great when the enterprise network was small. It became a
nightmare when the enterprise network became large. Building and
maintaining an enterprise network using layer three routing was a lot less
painful.
It is ironic that FDDI, which is an optical version of a token ring
technology has been successful. FDDI has continued to be the mainstay in
metro data services until the advent of GbE. Some RBOCs are still using
FDDI to provide 10mb transparent LAN services. Full duplex does GbE provide
an opportunity to upgrade to a higher bandwidth service.
Why did FDDI remain the dominant native data transport in the metro area for
so long? Why did full duplex 100BaseFX, with optical converters, not become
dominant? As far as I know, there is no technical reason why it could not
be used in the same way, or is there?
Until recently, FDDI was used, by at least one major computer system vendor,
as the protocol for its data storage farms, not full duplex 100BaseFX. Why
is that?
While I am suspicious of my figures, for a given bandwidth, I believe that
the optical interface for ATM is actually less expensive than the optical
interface for 10GbE. (Reminder, I wrote "the optical interface". ATM is
inherently more expensive because of the massive amount of data processing
required for segmentation and reassembly.) (I attempted to separate the
cost of the data processing from the cost of the optical interface to find
the cost of the optical signaling for different kinds of transport protocol
interfaces. If you can figure out a valid way of doing it, I would be glad
to look at it.) ATM OC rate uses SONET signaling, without the OAMP
processing, for the transport adaptation layer. For a given data traffic
rate, the SONET optical signaling only adds 4% overhead, while that used for
GbE adds 25% overhead. If the optical transport signaling for ATM is
actually less expensive for a given bandwidth than that used for GbE, what
does that say for what would be less expensive for 10GbE.
Thank you,
Roy Bynum
MCI WorldCom
Walter Thirion wrote:
> Au contraire... I strongly believe that ethernet has expanded to its
> dominant position because it is a relatively simple protocol. Having a
> background in token ring, which has obviously lost to ethernet, I
> learned a valuable lesson about simplicity. The ATM camp promised a lot
> of things because of superior managability, qos, etc. ATM was supposed
> to be extended to the desktop and replace ethernet. We all know what
> happened there...
>
> I find it ironic that many of the arguments put forth about ethernet in
> the WAN environment are very similar to the old arguments put forth by
> the token ring camp about the proper way to manage a LAN environment.
>
> Walt
>
> > -----Original Message-----
> > From: Roy Bynum [mailto:rabynum@xxxxxxxxxxx]
> > Sent: Wednesday, August 25, 1999 8:30 PM
> > To: Walter Thirion
> > Cc: HSSG
> > Subject: Re: AW: Deconstructing OAM&P
> >
> >
> > Walt,
> >
> > If you could turn back time, and do away with full duplex
> > optical 802.3, you
> > might could prevent the expansion of 802.3 into the WAN environment.
> > Unfortunately, time can not be erased, or the previous work
> > done by the
> > 802.3 WG. 802.3 is already being implemented over WAN
> > services. The only
> > problem is that it can not manage itself. I would much
> > rather see the HSSG
> > control the standard than allow other standards organizations
> > that have
> > other agenda's take control it.
> >
> > Thank you,
> > Roy Bynum
> > MCI WorldCom
> >
> >
> >
> >
> > Walter Thirion wrote:
> >
> > > This statement is exactly what many people are afraid of and what we
> > > tried desperately to separate in the speed ad hoc. Asking
> > 802.3 to buy
> > > into the WAN management protocol is going to be very
> > difficult. It is
> > > similar to the problem we've been having trying to separate the line
> > > code from the MAC/PLS speed.
> > >
> > > Walt
> > >
> > > > -----Original Message-----
> > > > From: Roy Bynum [mailto:rabynum@xxxxxxxxxxx]
> > > > Sent: Wednesday, August 25, 1999 8:17 AM
> > > > To: rtaborek@xxxxxxxxxxxxxxxx
> > > > Cc: Heiles Juergen; HSSG
> > > > Subject: Re: AW: Deconstructing OAM&P
> > > >
> > > >