Cell Relay Archive

["ronald h. davis" <ronaldd@lucent.com>]

albert.e.manfredi@boeing.com wrote:
> 
> In article <86zp1h52vz.fsf@ironbridgenetworks.com>,
>   James Carlson <carlson@ironbridgenetworks.com> wrote:
> > "O.B. Kwon" <obkwon@netcom.co.kr> writes:
> > > What's POS (Packet over SONET) ?
> >
> > Also known as PPP over SONET/SDH.  It's a way of transporting
> > datagrams over SONET/SDH channels.  See:
> >
> >       ftp://ftp.isi.edu/in-notes/rfc2615.txt
> >
> > > What's the different between POS and ATM ?
> >
> > POS uses variable-length datagrams as the basic unit of transfer.  ATM
> > uses fixed-size cells.  The rest, as they say, is marketing.
> >

there's also a matter of overhead.  when carrying packet over sonet
there's an intermediate layer: it can be ppp or atm.  as far as
overhead goes, ip/ppp/sonet is more efficient than ip/atm/sonet.
the value of marketing, in this case, would be to educate the customer
of the difference...


> > > Which technology is more advantage under multimedia environment ?
> >
> > ... and that's how you start a pointless flame war.
> 

flames follow below.  wear flame-retardant clothing...


> I agree with your last phrase, but not that the only difference beyond
> packet size is marketing. The differences are fundamental:
> 
> 1. Packets over SONET require the addition of layer 3 functionality in
> order for the system to achieve global reach. Packets over SONET is
> functionally identical to packets over T1, packets over T3, etc. SONET
> is a convenient pipe. You still need a network of routers along with
> all the trappings of the Internet to make such a  system useful.
> 
> 2. ATM over SONET, without any additional layer 3 protocol, is a system
> with global reach. It can be used as is to carry voice, video, or data
> (e.g. data can be sent using ATM's own error-free protocol, SSCOP). In
> other words, ATM over SONET was originally meant as an enhancement
> beyond ISDN. So it doesn't depend on IP or any other network layer
> protocol. Not by _any_ means to say this results in anything simpler
> than an Internet-type arrangement. But it is obviously a completely
> different beast.
> 

when you consider ip switches that can switch over oc-12 (and higher)
links at media speed, a network of ip switches interconnecting
ip over sonet trunks starts looking a lot like a network of atm
switches interconnecting atm over sonet trunks.


> If the only thing you want to do over the SONET is carry data frames,
> then POS is clearly a useful scheme. But if you want to use the SONET
> infrastructure for non-data services, ATM can give you that.
> 

		*** alert!!! potential flamewar bait ***

i agree here; it's when you want to carry traffic *other* than data
that atm begins to become a viable alternative.  when you compare
voice over ip (e.g. h.323) to voice over atm you can carry voice
traffic with greater protocol efficiency using atm.  furthermore,
with atm you are able to get guaranteed qos.

> It's also true that if all of the ongoing efforts in voice over IP, QoS
> over IP, etc. etc. pan out as is hoped, then one might question the
> need for having that ATM layer over SONET. After all, if IP can carry
> all of the data, voice, and video over a global network, no other
> scheme would be required.
> 

personally, i think that as the ip qos types tinker with protocols to
give better and better end-to-end qos guarantees they're going to find
out that qos isn't free...and in the end, to meet the demands of public
service providers they're ultimately going to end up with a "poor man's"
version of atm.

here's where you really get into marketing because the marketing types
are going to have to continually try to convince customers at every
step of the evolution that *this* time they really have gotten the
qos problem solved.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
author of "atm for public networks" published by mcgraw-hill
     http://www.amazon.com/exec/obidos/ASIN/0071344764