Path: utzoo!utgpu!watserv1!watmath!att!pacbell.com!ucsd!usc!samsung!munnari.oz.au!brolga!citrus.citr.uq.oz.au!gottsc
From: gottsc@citrus.citr.uq.oz.au (John Gottschalk)
Newsgroups: comp.protocols.iso
Subject: X.500 abandon and the ALS
Message-ID: <1991Jan31.015526.12318@brolga.cc.uq.oz.au>
Date: 31 Jan 91 01:55:26 GMT
Sender: news@brolga.cc.uq.oz.au
Reply-To: gottsc@citrus.citr.uq.oz.au (John Gottschalk)
Organization: CiTR, University of Queensland, Brisbane, Australia
Lines: 62
Nntp-Posting-Host: citrus.citr.uq.oz.au


 Hello everybody,
 
 
 I have been studying the Application Layer Structure (ALS) standard (ISO 9545)
 and have a concern about a possible conflict with the CCITT X.500 directory
 recommendations.
 
 
 The full set of X.500 DAP services are provided by three ASEs:

 --------------------------------------------------------------------- 
        ASE          |           service
 --------------------------------------------------------------------- 
                     |
        modifyASE:   |    AddEntry, RemoveEntry, ModifyEntry, ModifyRDN
        readASE:     |    Read, Compare, Abandon
        searchASE:   |    Search, List
 --------------------------------------------------------------------- 
                     
 
 Even though the "Abandon" service is part of the readASE, it may also be
 used to abandon an operation from the searchASE.
 
 
 In the definition of an ASE, the ALS standard states (5.5.1):
 
 "Grouping of functions into an ASE must contain at least all the
 functions and the corresponding APDUs which are required for a protocol
 machine which is logically complete and consistent in itself."
 
 
 This prompts the following question:
 
 As the protocol machines for the searchASE and readASE must be
 independent, how can it be possible for a searchASE to process an
 "Abandon" indication when its protocol machine does not contain this
 type of service?
 
 
 It seems to me that the searchASE and readASE are not really independent
 at all, and so violate the definition of an ASE in ISO 9454. If this is
 so then perhaps searchASE and readASE should be combined into one ASE.
 Note that if it is possible for one ASE to cancel an operation originating 
 from another ASE, then perhaps it is possible for the X.500 "Abandon" 
 operation to be applied to an ASE from another set of standards, 
 such as X.400 or network management, as long as all the ASEs are part of 
 the same application context. Perhaps ROSE should offer an "abandon" 
 service instead.
 
 Do other people think that there really is a conflict here, or have I
 interperted the standards incorrectly? Note that ISO 9545 was produced
 after the X.500 recommendations, which would explain any conflict, but
 any conflict needs to be resolved anyway.
 
 ===================================================================== 
     John Gottschalk (gottsc@citrus.uq.oz) 
     Center for Information Technology Research,
     University of Queensland, St Lucia, 4072, 
     Queensland, Australia,
     +61 7 365 4321 (phone), +61 7 371 3044 (fax)
 =====================================================================