Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: Does Ten-Gigabit Ethernet need fault tolerance?




Mick,

At 3:28 PM 99/7/16, Mick Seaman wrote:
>What needs to be built in is the detection of failure. What we don't need to
>do is to build everything into the MAC. I suggest you look at the fault
>tolerant capabilities provided by P802.3ad and the work on Rapid
>Reconfiguration starting in 802.1.

I agree that detection of failure is best built into the MAC hardware,
either in NICs or in hubs.  Actually, one needs to detect changes
(additions as well as losses) in network topology, so that nodes can rejoin
the network after a fault heals (such as a connector being plugged back
in).  In RTFC, what is detected and reported is the loss or gain of
modulated light at all receivers; any change triggers reexploration of the
network.

I haven't been following P802.3ad.  Could someone point me to the URL of
the relevant document?  Likewise Rapid Reconfiguration in 802.1.   How far
along are these efforts?  RTFC has been implemented.


>Both these (will) provide a degree of fault tolerance based on using
>protocols that are independent of MAC details to allow network nodes to
>precalculate their response to a low level indication of failure. There is
>really no need to build these protocols into the MAC.

It's true that given a topology-change signal from the MAC, one can
implement the rest of the algorithm elsewhere.  The advantage of
implementation in the MAC is that the vast existing base of software can
remain unchanged.  By the way, the rostering algorithm is quite simple, and
wholly independent of what goes on in "data mode" (normal LAN operation),
so it doesn't seem so onerous to put the rostering algorithm in the MAC.
So, it's a tradeoff.

It's also likely to be necessary in practice the implement all of the
algorithm in hardware, because pieces in software are vulnerable to
software problems.

If one implements the rostering algorithm in the hubs (not the NICs), the
NICs become very simple, and one can use the same NIC hardware for both
vanilla systems and for fault-tolerant systems; only the hubs would need to
change.  This would seem the best approach for 10GbE, as it would simplify
the overall system, and allow customers to reuse their plain-vanilla NICs
when upgrading their network to support fault tolerance.  Customers don't
like to throw perfectly good hardware away.


I'm sure a few brave souls, no doubt driven by a misplaced and dangerous
curiosity, will read the RTFC Principles of Operation, and will be in a
better position to assess where such an algorithm is best implemented, if
anywhere.   Aside from the necessity of keeping the software (including
especially I/O drivers) in the dark and out fo the picture, I really don't
have an opinion.

Joe


>-----Original Message-----
>From: owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx
>[mailto:owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx]On Behalf Of Joe Gwinn
>Sent: Friday, July 16, 1999 3:15 PM
>To: stds-802-3-hssg@xxxxxxxx
>Subject: Does Ten-Gigabit Ethernet need fault tolerance?
>
>
>
>The purpose of this note is to present a case for inclusion of fault
>tolerance in 10GbE, and to offer a suitable proven technology for
>consideration.  However, no salesman will call.

[snip]