The MPLS WG Archive

[neil.2.harrison@bt.com]

If you change a path then you change its delay.  In all cases except 1+1 (eg
1:1 or 1:N) you will get a hit on the restored path.....that's just a fact.
In 1+1 you have the ability, should you choose, to delay/phase align the
working and protection.  Clearly the working and protection path must be
buffered such that the 'faster' path has the same delay as the 'slower'
path......otherwise hitless changeover will not be posssible.

But you also have to ask yourself if a hit such a big deal?  This depends on
the applications tolerance to the hit it gets.........and this can vary and
is not simply a function of the transport delay, but what
framing/realignment process (if any) the application is using.

neil

> -----Original Message-----
> From: Pour, Mort [mailto:mpour@celoxnetworks.com]
> Sent: 28 June 2001 19:22
> To: 'mpls@uu.net'
> Subject: RE: 1+1 protection switching
> 
> 
> As far as I know there are no explicit requirements on reordering.
> 
> -----Original Message-----
> From: Saurabh Shrivastava [mailto:saurabh@timetra.com]
> Sent: Thursday, June 28, 2001 1:18 PM
> To: 'Pour, Mort'; 'mpls@uu.net'
> Subject: RE: 1+1 protection switching
> 
> 
> but isnt there a requirement that in an LSP, there
> should be no packet reordering.
> 
> 
> ~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~
> Saurabh Shrivastava          saurabh@timetra.com        
> 
> TiMetra, Inc.                408.245.8796 (residence)
> http://www.timetra.com/      650.237.5154 (office)
> ~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~
> 
> 
> *  -----Original Message-----
> *  From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On 
> Behalf Of Pour,
> *  Mort
> *  Sent: Thursday, June 28, 2001 11:13 AM
> *  To: 'mpls@uu.net'
> *  Subject: RE: 1+1 protection switching
> *  
> *  
> *  On second thought I agree. I don't think there will any 
> *  problem except for
> *  some
> *  retransmissions or some duplicate packets being dropped.
> *  
> *  
> *  -----Original Message-----
> *  From: Shahram Davari [mailto:Shahram_Davari@pmc-sierra.com]
> *  Sent: Thursday, June 28, 2001 10:49 AM
> *  To: 'Pour, Mort'; 'mpls@uu.net'
> *  Subject: RE: 1+1 protection switching
> *  
> *  
> *  Could you please explain more, why do you think that TCP 
> *  connection may be
> *  dropped.
> *  
> *  > -----Original Message-----
> *  > From: Pour, Mort [mailto:mpour@celoxnetworks.com]
> *  > Sent: Thursday, June 28, 2001 11:39 AM
> *  > To: 'mpls@uu.net'
> *  > Subject: RE: 1+1 protection switching
> *  > 
> *  > 
> *  > Thanks,
> *  > But this may also cause the connection to be dropped by TCP?
> *  > 
> *  > Regards,
> *  > Mort
> *  > 
> *  > 
> *  > -----Original Message-----
> *  > From: Shahram Davari [mailto:Shahram_Davari@pmc-sierra.com]
> *  > Sent: Thursday, June 28, 2001 10:36 AM
> *  > To: 'Pour, Mort'; 'mpls@uu.net'
> *  > Subject: RE: 1+1 protection switching
> *  > 
> *  > 
> *  > Hi,
> *  > 
> *  > I think it is understood/accepted that a number of packets 
> *  > may be dropped or
> *  > deliver out of order for a short period of time. The 
> *  receiver should
> *  > re-order the packets and/or request retransmission.  
> *  > 
> *  > Yours,
> *  > -Shahram
> *  > 
> *  > > -----Original Message-----
> *  > > From: Pour, Mort [mailto:mpour@celoxnetworks.com]
> *  > > Sent: Thursday, June 28, 2001 10:57 AM
> *  > > To: 'mpls@uu.net'
> *  > > Subject: 1+1 protection switching
> *  > > 
> *  > > 
> *  > > 
> *  > > I have a question regarding the 1+1 protection switching 
> *  in MPLS. 
> *  > > Draft <draft-ietf-mpls-recovery-frmwrk-02.txt>
> *  > > states that the path merge LSR can select the backup path 
> *  > > when the primary
> *  > > path fails.
> *  > > What about the issues of duplicate (or missing) packets on 
> *  > > the backup path
> *  > > due
> *  > > to variations in delay between the two path. Or if the 
> *  two path are
> *  > > synchronized, how is that achieved?
> *  > > 
> *  > > Thanks
> *  > > Mort Pour
> *  > > 
> *  > 
>