[15804] in North American Network Operators' Group
Re: IP over SONET considered harmful?
daemon@ATHENA.MIT.EDU (Vadim Antonov)
Sun Mar 22 01:17:29 1998
Date: Sat, 21 Mar 1998 14:04:05 -0800 (PST)
From: Vadim Antonov <avg@pluris.com>
To: avg@pluris.com, smd@clock.org, stan@networkmercenaries.com
Cc: hannigan@xcom.net, nanog@merit.edu
Sean M. Doran <smd@clock.org> wrote:
>| Native IP over fiber is not a religion; it's a solid engineering.
> I'm not totally convinced of this yet, because of the useful
> O&M and restoration facilities SDH/SONET buys you now.
There's LQM in PPP.
> You need something underneath IP to do the sort of Lothberg
> restorations I think you saw explained at the last NANOG meeting.
Well, but it doesn't have to be synchronous and it does not
have to involve any additional baroque framing. Better yet,
what i do with splicing traffic into many channels effectively
provides IP-invisible restoration w/o loss of useful capacity.
Just route those channels into diverse physical paths :)
> SDH/SONET is also a really keen way of sharing a network
> among IP and other things on the wide-area side, and
> an even keener way of having access to relatively low-speed
> customer data on the more local side.
On customer-access side even ATM seems to be fine :) It
definitely is a huge lot cheaper than SONET gear.
As for sharing fiber between different networks WDM seems to be
just fine - as soon as IP network gets to take an entire OC-48
slice.
> I agree with you that there is alot of unnecessary overhead
> in SDH/SONET, but there is also alot of what seems now to
> be necessary overhead, too.
I would venture to state that _all_ SONET/SDH overhead is unnecessary.
Getting rid of SONET framing and delivering bit stream from WDM muxes
to routers allows to make WDM muxes a whole lot simpler and cheaper -
by nearly eliminating electronics in the data path. Ciena folks told
me some other compelling reasons for dropping SONET, which i do not
feel i'm at liberty to discuss here.
>I do, however, expect heavy
>experimentation involving radical simplification of on-fibre
>framing for point-to-point paths dedicated exclusively to IP.
I also expect it :) My current favourite for framing is
32-to-33 bit encoding, with flag being one of "malformed words",
one word header (2 bytes for payload length, 2 bytes for tag), and
32-bit CRC at the end. I.e. the per-packet overhead is 12 bytes + 3.1% +
rounding to 4 bytes for odd-sized packets. Make all 1s and all 0s
and chess patterns to be invalid words, and loss-of-carrier becomes
easy to detect.
--vadim
