Session was chaired by Volker Sander, Research Centre Jülich 
- IP statement presented 
- Jim Pruyne replaces Jon MacLaren as co-chair -> accepted by the group Web page    
   will move; Jon MacLaren will add a redirect. Thanks to Jon who was a driving force    
   of GRAAP as chair, we are looking forward continuing the collaboration 
- Consensus on the proposed Agenda 
- GRAAP milestones re-confirmed 
- Charter: clarifications of the scope -> no objections, consensus on deleting the   
   clarification i.e. "do you also include the reservation of ports for communication, or 
   allow the protocol to re-configure firewalls?" have been removed). Agreed on 
   proposal to add the other changes not only to the website but to add it to the charter  
   and sending the updated charter to the steering committee for approval (a formal  
   issue mostly as the clarification does not change charter, but gives more more focus 
   to it) 
- Other groups to link to: 
   o    proposed CRM WG 
   o    proposed JSDL WG  (Jim Pruyne) 
   o    proposed GSO WG (Philipp Wieder) 
   o    Other links contact via email 
- Definition of AR -> no objections on the definition. GRAAP will use this definition 
- State-of-the-art document -> no objection, no new add-ons. We will pursue to move 
   this document towards an GWD-I document 
- Scenarios 
   o   The Templated Application: proposal to change the protocol requirements section   
        with respect to the issue of composition of existing SLAs -> the term SLA might 
        be misleading here. 
        Decision: leave it for now but clarify  and decide later 
   o   Discussion about the modification of reservation parameters: totally cancelling a 
        reservation or reducing the number (nodes, bandwidth, disk-space,...) of reserved 
        resources -> does it have to succeed? 
        Is it a protocol issue, a policy issue, or a recommendation how to use the 
        protocol? 
        -> agreement: re-negotiation must be supported by the protocol, semantics of   
        reducing or cancelling a reservation does not have to be specified within the 
        requirements. The common understanding is that this seems to be part of the SLA 
        policy. 
- Requirements 
   o  Reducing the SLA should succeed will be changed to a less obligating 
       formulation as one might not be able to enforce what happens at the other 
       side, but an acknowledgement that the request was received and will be       
       processed somehow should be returned (see to discussion above)  
   o  Any requirements missing? 
         * Use of GRAAP for co-allocation: Two phase commit/soft-state model? 
              -> requirement: as SLAs may have different life-times the need for             
              validations might come up at different times. Add a co-allocation                         
              scenario to the document. (for two-phase commit/soft-state issues). 
         * Q: Is the protocol centred approach compliant to OGSA?
            A: GRAAP will be coded in WSDL with descriptions for state changes,          
            etc. -> full protocol 
- Should GRAAP address requirements for an RDL? 
   o  Q: Write an additional document specifying GRAAP s requirements           
       concerning an RDL (plus adding a milestone to the group's charter)? 
          -> Comment : Requirements specification for RDL may be beyond the 
             scope of the group. 
          -> agreement: specification of the protocol first and extraction of 
             the RDL requirements later from the protocol, no separate document  now. 
- All goals defined for GGF7 have been reached.