The MPLS WG Archive

[Dave McDysan <dave.mcdysan@wcom.com>]

Neil, All,

I propose that the we try to agree on the concept of a supporting a
"heartbeat" for important tunnels for the purpose of fault detection is an
important requirement.

In the mplsoam bof I offered to provide constructive feedback on the
requirements document. I suggest that we start trying to state what we agree
on as a requirements before evaulating solutions.

I may try to call you next year (vacation coming soon) to discuss ATM OAM,
in particular CV. My biggest concern with it was that in ITU (and ATM Forum)
it is optional for VCs, and was never widely implemented by vendors.

Regards,

Dave

> -----Original Message-----
> From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On Behalf Of
> neil.2.harrison@bt.com
> Sent: Thursday, December 13, 2001 7:37 PM
> To: dave.mcdysan@wcom.com; giles@packetexchange.net;
> Shahram_Davari@pmc-sierra.com
> Cc: dwfedyk@nortelnetworks.com; Ben.Mack-Crane@tellabs.com; mpls@UU.NET
> Subject: RE: MPLSOAM BOF meeting draft minutes
>
>
> Dave McDysan 13 December 2001 18:37
> > > -----Original Message-----
> > > From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On Behalf Of Giles
> > > Heron
> > > Sent: Thursday, December 13, 2001 1:20 PM
> > > To: Shahram Davari
> > > Cc: Don Fedyk; Ben Mack-Crane; neil.2.harrison@bt.com; mpls@UU.NET
> > > Subject: Re: MPLSOAM BOF meeting draft minutes
> > >
> > >
> > >
> > > I don't measure every circuit.  And what I do measure I measure for
> > > other purposes than "OAM".  It has a side effect of also
> > allowing me to
> > > do CV at a core MPLS tunnel level.
> > >
> > > I'm hoping that I won't see two blackholes a month.  And I
> > stand by my
> > > statement that if I did I'd be beating on my LSR vendor pretty darn
> > > hard!
> > >
> >
> > In the end, I believe that this is the real solution to this
> > problem. I
> > haven't yet seen discussion on this thread about what would
> > be done if an
> > MPLS OAM CV with BDI did detect a fault. Do we reboot the
> > entire node? What
> > if it is a software/hardware bug that affects multiple nodes
> > from a certain
> > vendor?
> >
> > Dave
> NH=> Fair point Dave.  You are right this has not indeed been decided
> upon....but could we even attempt to do that (I mean standardise
> what needs
> doing to the node....it depends what we are seeing)?  At least
> the 1st step
> would be to detect, and diagnose, the problem....and taking local
> executive
> actions to protect traffic as needed, eg protection switching.
> When you get
> down to it what is CV? It's merely a heartbeat that is decoupled from user
> activity saying source X is correctly connected to sink Y.....and it beats
> me why this functionality (so often encountered elsewhere...perhaps we
> should call it a 'hello' packet?) seems to create such a problem for some
> folks.
> BTW - I saw your earlier note on ATM OAM and if you want to discuss this
> I'll willingly do it off-line...or give me a call (+44 1 604 820
> 724).  ATM
> OAM has some subtle problems.
>
> regards, Neil
>