Re: LCUP: convergence?

["Timothy Hahn" <hahnt@xxxxxxxxxx>]

Hi all,

I agree with Kurt on this.

I believe that the LCUP draft needs to clearly state that "eventual
convergence" is the goal and show how the protocol can, indeed, be used to
get there (through appropriate use of synchronize, or synchronizeAndPersist
operations, for example).  Right now, it is unclear how to use the protocol
to get "eventual convergence" in a client.

Also to be covered should be a discussion of when it can be assumed by the
client to be "safe" to use only "persistOnly" and still be sure of gaining
"eventual convergence".

Equally important is a discussion of how clients will know that, with the
request that was made, "eventual convergence" cannot be ensured by the
server. (so that a client can re-formulate its held data and LCUP
request(s).

Regards,
Tim Hahn

Internet: hahnt@xxxxxxxxxx
Internal: Timothy Hahn/Durham/IBM@IBMUS
phone: 919.224.1565     tie-line: 8/687.1565
fax: 919.224.2540



                                                                                                                                       
                      "Kurt D.                                                                                                         
                      Zeilenga"                To:       <ietf-ldup@xxxxxxx>                                                           
                      <Kurt@xxxxxxxxxxx        cc:                                                                                     
                      g>                       Subject:  LCUP: convergence?                                                            
                      Sent by:                                                                                                         
                      owner-ietf-ldup@m                                                                                                
                      ail.imc.org                                                                                                      
                                                                                                                                       
                                                                                                                                       
                      09/11/2002 01:18                                                                                                 
                      PM                                                                                                               
                                                                                                                                       
                                                                                                                                       




In my analysis of the various possible uses of LCUP, I
see a strong requirement for the protocol to provide
"eventual convergence" of the synchronized content.
However, in reviewing the LCUP I-D (see previously
posts for details), I see that it provides for only
"lossy convergence".  That is, LCUP I-D allows the
synchronized content to diverge.  It is not clear to
me that "lossy convergence" meets the needs of the
applications described in the document.

What's convergence level(s) are intended to be
supported by this protocol?

Kurt