SSSWG - September 1996 Minutes

IEEE Storage Systems Standards Working Group (Project 1244)

September 15-16, 1996 Minutes

Meeting Held at College Park, MD
in Conjunction with the 5th NASA/Goddard Storage Conference

Chair:  Jack Cole, Minutes:  Linda Kempster & Jack Cole
Documents Custodian: Dixon Hutchinson
Archivist:  Bill Wolfe, NASA/Langley Research Center

Attendance:

	Jack Cole, Army Research Lab
	cole@arl.mil
	
	Sam Coleman, Lawrence Livermore Lab
	scoleman@llnl.gov
	
	Dixon Hutchinson, Exabyte Corp
	hutch@exabyte.com
	
	Merritt Jones, The MITRE Corp
	merritt@mitre.org
	
	Linda Kempster, IIT Research Institute
	lkempster@mtc.iitri.com
	
	Eric Stouffer, IBM/Tucson
	stouffer@vnet.ibm.com
	
	Lee Ward, Advanced Engineering Research Assoc.
	lee@ava.com

	Joel Williams, Systems Engineering and Security
	joel.williams@ses-inc.com


10:30 am Sunday, September 15, 1996 through 5 pm Monday September 16, 1996

   This special meeting of the SSSWG was begun on Sunday in order to fit
in at the beginning of the 5th NASA/Goddard Mass Storage Systems and
Technologies Conference.  Ordinarily meetings are held Tues- Thursday,
and it is unlikely that the SSSWG will ever meet again on a weekend.
Nevertheless, eight dedicated individuals attended!  

   This meeting marks the beginning of an effort to accommodate
participation by non-attendees in view of the fact that travel budgets
for many interested parties have acted to prevent their attendance.
That is, there are many individuals with a strong desire to participate
in the development of storage systems Model and standards, but unable to
attend many meetings.

   Also remarkable at this meeting was the beginning discussion of
revision of Version 5 of the Model, to complete hopefully within a year
of this beginning.  Sam Coleman has indicated that he will serve as the 
Version 6 Editor, especially if we can devise a way to accommodate
infrequent attendees.  Merritt is positive that we can accomplish this.

   Attendance was taken, roster information was updated, and
registration completed.  Minutes from the Breckenridge meeting in July
were read and approved, as was the agenda for this meeting.

   Because of the impending discussion of the Model, a review of the
history of this WG and the Model was lead by Merritt and Sam.  Merritt
commented that anyone can bring in her or his flavor of the APIs being 
developed for standards by this group.

   Joel presented local restaurant information and slides of the status
of PVR and PVL.  The PVR API is on the WEB and close to being a
ballotable document.

   Dixon described the current status of the MCS PAR, and it appears
that little remains to complete this document for submission by the
November 1 deadline.  "PAR" refers to the IEEE Project Authorization 
Request process to begin work on a new standard.  And "MCS" is the new
standard being considered --- the Medium Changer Service Standard for
robotics interface based on the functionality of the SCSI command 
set.

   The discussion of revision of the Model began with Eric remarking
that the Model should reflect how the world works today.  That is, it
should be a roadmap to the real world.  This remark sparked a discussion
of 'real world' storage system problems, such as transitioning data from
one HSM to another.

   Concern was expressed that the Model revision effort would distract
from the effort to get PVR and PVL to ballot.  Merritt countered that
the effort would result in stronger APIs when inconsistencies between
the PVR, PVL and the Model were removed.  

   Discussion of the PVR status and the Model revision lead to a third
discussion on the scope of the management (MGT) and virtual storage
server (VSS) components of the Model.  While HSMs fit within the 
VSS component, the group felt that a complete VSS API would be too
difficult to pursue because of the complex functionality. For example,
it  is this component that addresses metadata issues for the Model.  
After a long discussion of management functionality, it was concluded
that management functionality could reside entirely within each of the
Model components (such as PVR) or elsewhere.  And that the two blocks 
on the right side of the drawing on page 26 of the Model could be
combined.

   Joel began a discussion of the relationship of DCE to the Model and
its components.  This discussion was similar to the one held in July
1995 at the DEC facility in Bellevue, WA.  That is, will the Model
force implementors to use DCE or any other transfer agent.  The answer
again was "no", but it was agreed that the Model could state that we
believe DCE provides everything needed by the AEP, although DCE itself
is not necessary if this functionality can be provided otherwise.

   The meeting ended with an exposition by Eric on IBM's ADSM.


Action Items:

Eric - Provide ADMS device driver interface information as input to the
MCS effort.

Dixon - Arrange an exploratory meeting among vendors to discuss MCS.

Jack - Get the PAR out, get the Model into HTML format, and work out the
details of non-attendee participation.

Sam - Arrange the Timberline Lodge meeting, serve as Version 6 Editor.

Linda - Coordinate the concurrent THIC-SSSWG meeting in January.


Next Meeting:

November 5-7, 1996   Cupertino, CA   HP/ Hamid Vazire
(Meeting preceded by a vendor only meeting on Monday November 4th arranged
(by Dixon Hutchinson to discuss MCS, PVR, PVL)
(attendees need to obtain absentee ballot for General Election on Nov 5)


Tentative Schedule of Future Meetings:

January 22-24, 1997   Seattle, WA   Hilton/ Linda Kempster
(Meeting arranged to overlap THIC meeting same location,
Linda and Jack coordinating)

March 11-13, 1997   Colorado   Location?/ Dixon Hutchinson
	
May 13-15, 1997   Key West, FL   hotel/ Jack

July 15-17, 1997   Colorado   location? /Dixon

September 16-18, 1997   Portland, OR  Timberline Lodge/ Sam