Re: [mpls-tp] 答复: Re: Alarm Reporting (aka AIS)

[<[e-mail]>]

Huub wrote 13 December 2010 10:21

> Hello Steven,
> 
> You wrote:
> 
> > I think we need to listen to the views of Neil on AIS.
> >
> > This is just a purely technical discussion, isn't it?
> 
> Yes it is, I am *not* trying to shoot the messenger.
> 
> My concern is that we have a requirement in RFC5860.
> A solution (for AIS) is provided in draft-ietf-mpls-tp-fault.
> 
> How will the current discussion affect the progress of this draft?

NH=> I guess a key issue here is trying to figure out which, if any, operators still want AIS.  Perhaps those operators did not previously understand that:
-	is it a redundant technical function (for reasons I have explained previously), and 
-	it is a potential source of further problems in itself, as OAM messages have to be constructed/configured (the always-on defect detection needs to be as simple/sparse as possible since it must be significantly more reliable than the network it is monitoring, and thus all unnecessary sources of defects/configuration-errors should be minimised).

Perhaps those operators who may have originally supported the requirement for AIS in packet-switched networks did not properly understood these issues before?

In any case, it is important that for those operators who have properly understood what AIS is to be able to have this function disabled in their networks.

regards, Neil

> 
> Regards, Huub.
>