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

RE: [EFM] Active to active connections are useful

Tom - 

OAM has two provisioned modes - active and passive.  Active is intended for those devices that wish to query/control their peer in some respect.  Passive is intended for those that do not wish to query/control, but will respond.

OAM can be provisioned to be symmetric (e.g. active-active as John mentioned), or not (active-passive).  Both deployment models (symmetric, asymmetric) have been requested and determiend to have merit, and both can be supported.  

- Matt

> -----Original Message-----
> From: Tom Mathey, Gail McCoy []
> Sent: Thursday, September 18, 2003 6:40 AM
> To:; John Messenger; Alan Weissberger
> Subject: Re: [EFM] Active to active connections are useful
> John,
> I must admit to not following the Clause 57 OAM very closely. 
> However, from the reading that I have done, I certainly was under the 
> assumption that the OAM capability was symmetric.  That is, the 
> customer can perform any and all OAM actions.  The CPE end was no 
> different from the CO end.  A customer must be able to test their 
> connectivity to the CO.  If this is not permitted, then the draft 
> must change.  Do you need any support on this for the next draft?
> You must not be at the Italy meeting.
> A copy to Alan as he is interested in OAM.
> NOTE:  You need to modify your e-mail settings as the private message 
> that I tried to send on this same subject bounced with the following 
> reply:
>     ----- The following addresses had permanent fatal errors -----
> <>
>      (reason: 550 5.7.1 This system has been configured to 
> reject your mail)
> Tom
> >Hi,
> >
> >I read the comments about "local_satisfied" in the proposed 
> resolutions
> >(#594, #679) and I agree that it would be helpful to include 
> diagnostics
> >about why discovery won't complete.  However it raises a 
> couple of issues:
> >
> >1. The ability to initiate a loopback from the CPE to the CO 
> is useful for
> >commissioning.  This requires the CPE to be in Active mode.  
> (Just because
> >you let the other end be Active doesn't mean you have to support his
> >Variable Requests.)  I'd be interested to know whether 
> switch vendors intend
> >to support OAM loopback paths for this purpose?
> >
> >2. I anticipate implementations that propose their favoured 
> configuration
> >and see if it is acceptable to the other end, but if it is not
> >(local_satisfed remains false after a few cycles of 
> discovery) then back off
> >to a less-favoured but still operable configuration.  For 
> example a CPE
> >might propose to be active and if that was not OK, back off 
> to proposing to
> >be passive.  As we do not specify the behaviour of the OAM 
> Client, this is a
> >point that implementors need to bear in mind - discovery should be a
> >negotiation, not a "take it or leave it" conversation.
> >
> >If anyone would be willing to summarise the resolution at 
> the interim for
> >me, I'd be grateful.
> >
> >	-- John
> >--
> >John Messenger (
> >R&D Software Manager, ADVA Optical Networking Ltd.
> >Tel: +44-1904 699309   Fax: +44-1904 699378
> >
> >"Mind like parachute - only work when open" (Charlie Chan)
> --