SSSWG - July 1995 Minutes
IEEE Storage System Standards Working Group Meeting
18-21st July, 1995
Digital Equipment Corporation, Seattle /
Bellevue, Washington.
Attendance:
Carl Braganza (Advanced Computing Systems
Company, Inc.)
Tim Boyce (IBM, Austin)
Jack Cole (Army Research Laboratory)
Anat Gafni (Advanced Computing Systems
Company, Inc.)
Richard Golding (Hewlitt-Packard Laboratories,
Palo Alto)
Bruce K. Haddon (Storage Technology European
Operations)
Merritt Jones (MITRE Corporation)
Michael T. Peterson (Digital Equipment
Corporation, West)
Dave Skinner (Fujitsu)
Eric Stouffler (IBM Storage Systems Division,
Software Products)
Hamid Vazire (Hewlitt-Packard, Cupertino)
Rich Wrenn (Digital Equipment Corporation,
Co. Springs)
Joel Williams (MITRE Corporation)
Chair: Dave Skinner.
Recording Secretary: Bruce K. Haddon.
General Session, Tuesday 18th July, 9.00 a.m.
Agenda
======
(This agenda reflects minor changes made in the course of
the meeting.)
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of action items from previous meeting
Status of recruitment and sponsorship.
Status of issued RFP's
Posix Update
Discussion of "session" abstraction (Dave E.
Skinner)
Subcommittee Sessions
Discussion of "non-cooperating clients and
security" (Bruce K. Haddon)
Subcommittee Sessions
Discussion of "base types" (Mile Peterson)
Closing General Session
PVL subcommittee report(Mike Peterson)
PVR subcommittee report(Dave Skinner)
MVR subcommittee report(Bruce K. Haddon)
Tabled items
Review of open action items
Dates
Agenda for 95/9 meeting
1. Announcements
=============
"Early registration" for MSS (closes mid-August, please
attend)
Local Seattle meeting details.
Copies available of:
PVR Draft
MVR Draft
"The Role of Standards in Storage Solutions"
presented at Network Storage 95, by Dave
Skinner.
2. Minutes of previous meeting
===========================
It was decided at the May meeting that minutes are to
carry organizational affiliations (and should have this
last time). This will be implemented in the minutes of
the current meeting.
## Resolved: that the minutes of the previous
meeting be approved.
Moved: Mike Peterson; 2nd: Joel Williams.
Carried without dissent.
3. Treasurer's report
==================
Copies of the balance sheet were handed out and
presented. Copies of the presented balance sheets are
available from Dave Skinner. Jack Cole will receive dues
for this meeting.
4. Review of action items from previous meeting
============================================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
Status: No action has been taken. Due
date is to be left unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: 5/31.
Status: Time does not permit
considering this RFP at this time.
Due date to be changed to
unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Creig's status continues to be
uncertain. There will be no way to
react to agreement on naming
conventions before September, hence
this item will be left with an
unspecified due date.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
Status: This matter was discussed at
the Colorado Springs meeting (March,
1995). No consensus has emerged so
the need for the discussion that the
RFP would elicit is probably still
needed.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Status: A wording for this RFP was
proposed by Rich Wrenn at the last
(May, 1995) meeting. This was re-
presented and a couple of
modifications proposed. This item
will be taken over by the PVR group
(Chair: Dave Skinner), and Rich will
submit the RFP text to the reflector
on behalf of the PVR group by the end
of July.
6. Status of the "draft" if the
Reference Model (must it be renewed,
or is it a "guide" and does not need
to be renewed).
Responsible: Merritt Jones.
Due: 95/7 meeting.
Status: Merritt has submitted a
request to the Technical Committee
asking that the Reference Model
document be recognized as a "guide",
and that it be exempted (permanently)
from the expiry clause. This has been
passed on to the IEEE for voting at
the next scheduled meeting. Status of
this request will be reviewed at the
95/9 meeting
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
Status: No change.
5. Status of recruitment and sponsorship
=====================================
As anticipated at last meeting, Dave Skinner presented a
paper at Network Storage 95 (a conference organized by
the other Mike Peterson) brought about half-a-dozen
inquiries to Dave Skinner and Rich Wrenn about the
function of the SSSWG and the role of the Mass Storage
Symposium. Mr. Peterson's organization collects industry
intelligence, focussing mostly on the "industry sweet
spot", rather than high end.
A need was expressed for brochure's etc., for the MSS at
Monterey in order to elevate level and interest and
increase exposure of the work of the SSSWG. No action
taken.
6. Status of issued and open RFP's
===============================
OID
Issued: Bob Baird (VSS)
It was presented at the last meeting that there was one
response and some subsequent discussion. Bob Baird has
updated the response and planned a final report to be put
on the ftp site and/or WEB page. It is not known whether
this has been done.
Base Types
Issued: Mike Peterson (PVL)
Two responses: To be added to the agenda for the next
meeting.
Criteria: consistency requirements between model
components.
Issued: Mike Peterson (PVL)
7. Posix status
============
Dave Skinner met with "Posix.1k" last week. They are
interested in an interface standard to augment or extend
the existing SCSI "robotics" interface. Dave suggests
that this is an appropriate piece of work for P1244. A
new PAR would be required, and would require additional
people interested in this level of work.
Hamid Vazire placed this work in context with what is
currently being done in P1244: The Posix group is trying
to standardize the "mtio" interface for serial access
media. Functional entry points are being defined to take
over operations that are now done through (largely non-
standardized) IOCTL functions. This work does not impact
P1244.
There was discussion of whether there was existing
overlap between the "media changer" interface and PVR
functions. There were views of both sides of this
question, but it was agreed a "clean" layering could be
defined.
There was then a rather complicated discussion about the
layering of OS specific interfaces (such as the Posix
interface) with 1244 interfaces. Bruce and Rich finally
agreed that layering could be done in both directions.
Bruce was volunteered to provide a set of pictures and a
summary of this discussion.
Dave will carry the message to Posix.1k that we support
the work on "mtio.h" and the "mt" utility. He will
continue to recommend that the media changer work should
be done in 1244, but it is not expected that this will
lead to any action at this time.
8. Discussion of "session" abstraction.
====================================
Dave re-presented the content of the paper that he had
earlier distributed via the reflector.
Rich pointed out that "allocate" means to "assign for
exclusive use". "Deallocate" is the release of this
condition. "Reservations" are intended to be essentially
notifications of coming allocations.
Carl described the "ALLOCATE" operation proposed for the
PVL, which is essentially an atomic operation on a set of
resources. It was pointed out that this is essentially a
"reserve and allocate" in one operation.
There was then a discussion of optimization and over-
commitment algorithms, and the need to accommodate legacy
systems, all of which argued in favor of the two step
model. The PVL group said that they would consider these
matters.
Rich was volunteered to summarize the discussion by
preparing material that would address:
1. Updating Dave's summary of the
"session" abstraction to have a
separate "reserve" and "allocate"
operation (rather than the collective
"commit" operation that causes an
"allocate" on every current
reservation.
2. Similarly, an update to the diagram
in Dave's email to represent the
change defined in (1).
3. To draft a statement that defines the
applicability of this updated session
model to each of the components of
the 1244 model
The session adjourned at 12:01 PM, the general session to
be resumed Wednesday morning.
General Session, Wednesday, 19 July, 9.02 am.
9. Discussion of "non-cooperating clients and security".
=====================================================
Bruce K. Haddon re-presented the argument that he had
earlier distributed in his working paper, entitled "Non-
co-operating clients, security, and related issues". The
essential idea was that the requirement for non-co-
operating clients was real and on-going, and that a
"security" methodology could provide a means of achieving
that requirement.
In the course of discussion, it became evident that the
term "non-co-operating" was unfortunate. This term is
used in operating system theory in the context of "co-
operating sequential processes", which relates to the use
of various mechanisms for synchronizing access to shared
resources, and, as such, is a component of the resource
allocation and deadlock avoidance scenario. In the
context of the OSSI model, the term refers simply to
clients that are using the same services, but are not in
communication with each other concerning those services.
The term "mutually unaware" was offered as a better way
to describe this scenario.
It also became apparent that it is the "access control"
aspects of security that are the pieces needed to effect
the creation of an environment in which clients could use
a service and be mutually unaware. So the working paper
presented by Bruce may have been better titled "Mutually
unaware clients, access control, and related issues".
A final agreement was not reached on the issues raised by
Bruce in the working paper. The final position is
summarized as follows:
1. There shall be a model developed for
access control that can be adopted by
each of the 1244 components.
2. Access control is recognized as being
sufficient to satisfy what was termed
in the working paper as the PVL
Requirement, viz.: "The PVL must be
able to guarantee deadlock free
mounts (of cartridges) to the PVL's
clients".
3. There still exists disagreement as to
whether the PVR should be permitted
to have mutually unaware clients.
The session adjourned at 11:44 am, the general session to
be resumed Thursday morning.
General Session, Thursday, 20 July, 8.59 am.
The discussion of "base types" scheduled at this time was
postponed to a time to be determined. Bruce was again
volunteered to update his earlier submission to the
reflector to incorporate more recent thinking.
The session adjourned at 9.15 am, the general session to
be resumed Thursday afternoon.
10. PVL subcommittee report(Mike Peterson)
======================================
Mike reported that the PVL subcommittee had met on the
Monday preceding this meeting, or order to make
additional progress on their draft API, and now, during
the meeting week, had been and were continuing to spend
the major proportion of its time on writing down the
various items upon which the committee had agreed.
He felt that the PVL API had solved some major problems,
but in terms of writing effort, still had a long way to
go, and that it was the PVL subcommittee's intention to
continue working on this at the end of the general
session and on into tomorrow
11. PVR subcommittee report(Dave Skinner)
=====================================
Dave reported that the draft API was in an advanced state
of writing, had been transformed into the desired "man"
page format, and was now being subjected to some overall
organizational corrections.
The main piece to add, which had been the subject of
current discussion, was concerned with security, and the
security interfaces. A proposal would be made for text
covering this subject that would be general in nature, as
there were still issues to be resolved with interactions
with the PVL.
There could be a chance, if it were possible for the PVR
subcommittee to get together between now and September,
for a draft that could be handed out at the MSS.
12. MVR subcommittee report(Bruce K. Haddon)
========================================
Bruce reported on behalf of Creig Humes. A draft of the
MVR API was available, also in the desired "man" page
format, with introductory sections concerning the mover's
objects, and their characteristics. In many ways, the MVR
draft was in the best shape of the three.
However, the question of security, asynchronous events,
and inter-MVR protocols were yet to be addressed.
13. Tabled items
============
None.
14. Review of open action items
======================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Rich Wrenn for Dave
Skinner (MVR).
Due: 7/31.
6. Update of status request for "guide
status" of the draft of the Reference
Model.
Responsible: Merritt Jones.
Due: 95/9 meeting.
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
8. Diagrams showing multiple layering of
MVR's and device drivers and MVR's,
etc.
Responsible: Bruce K. Haddon
Due: unspecified.
9. Summary of issues and refinement of
"session" concept.
Responsible: Rich Wrenn.
Due: unspecified.
10. Re-issue of data representation
response by Bruce.
Responsible: Bruce K. Haddon
Due: 95.9 meeting.
15. Dates
=====
* Dates for 1995
May 16-19: Valley Forge, Pennsylvania. (Host:
Rich Garrison)
July 18-21: DEC, Seattle. (Host: Michael
Peterson)
Sept. 14-15 (Monterey, at the end of the MSS)
Nov. 7-10 (ACSC, Los Angeles, Anit Gafni)
* Dates for 1996 (partial list)
Jan 16-19, Tucson, Arizona (Host: Eric
Stouffler, IBM).
Mar 12-15 (tentative, Colorado)
June (MSS is in Annecy, France). (Note added
later: the 1996 MSS planned for
Annecy in June has been postponed. A
new date is not yet available.)
16. Proposed agenda for 95/9 meeting
================================
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of Action Items from previous meeting
Status of recruitment and sponsorship.
Status of RFP's.
Subcommittee Sessions
Closing General Session
Subcommittee status (each Chairman)
Tabled Items
Review of Action Items
Agenda for 95/11 meeting.
## Resolved: that this proposed agenda be
accepted.
Moved: from the Chair.
Passed unanimously.
17. Other business
==============
None.
Session was adjourned 4 p.m., Thursday, 20 May, 1995.
IEEE Storage System Standards Working Group Meeting
18-21st July, 1995
Digital Equipment Corporation, Seattle /
Bellevue, Washington.
Attendance:
Carl Braganza (Advanced Computing Systems
Company, Inc.)
Tim Boyce (IBM, Austin)
Jack Cole (Army Research Laboratory)
Anat Gafni (Advanced Computing Systems
Company, Inc.)
Richard Golding (Hewlitt-Packard Laboratories,
Palo Alto)
Bruce K. Haddon (Storage Technology European
Operations)
Merritt Jones (MITRE Corporation)
Michael T. Peterson (Digital Equipment
Corporation, West)
Dave Skinner (Fujitsu)
Eric Stouffler (IBM Storage Systems Division,
Software Products)
Hamid Vazire (Hewlitt-Packard, Cupertino)
Rich Wrenn (Digital Equipment Corporation,
Co. Springs)
Joel Williams (MITRE Corporation)
Chair: Dave Skinner.
Recording Secretary: Bruce K. Haddon.
General Session, Tuesday 18th July, 9.00 a.m.
Agenda
======
(This agenda reflects minor changes made in the course of
the meeting.)
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of action items from previous meeting
Status of recruitment and sponsorship.
Status of issued RFP's
Posix Update
Discussion of "session" abstraction (Dave E.
Skinner)
Subcommittee Sessions
Discussion of "non-cooperating clients and
security" (Bruce K. Haddon)
Subcommittee Sessions
Discussion of "base types" (Mile Peterson)
Closing General Session
PVL subcommittee report(Mike Peterson)
PVR subcommittee report(Dave Skinner)
MVR subcommittee report(Bruce K. Haddon)
Tabled items
Review of open action items
Dates
Agenda for 95/9 meeting
1. Announcements
=============
"Early registration" for MSS (closes mid-August, please
attend)
Local Seattle meeting details.
Copies available of:
PVR Draft
MVR Draft
"The Role of Standards in Storage Solutions"
presented at Network Storage 95, by Dave
Skinner.
2. Minutes of previous meeting
===========================
It was decided at the May meeting that minutes are to
carry organizational affiliations (and should have this
last time). This will be implemented in the minutes of
the current meeting.
## Resolved: that the minutes of the previous
meeting be approved.
Moved: Mike Peterson; 2nd: Joel Williams.
Carried without dissent.
3. Treasurer's report
==================
Copies of the balance sheet were handed out and
presented. Copies of the presented balance sheets are
available from Dave Skinner. Jack Cole will receive dues
for this meeting.
4. Review of action items from previous meeting
============================================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
Status: No action has been taken. Due
date is to be left unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: 5/31.
Status: Time does not permit
considering this RFP at this time.
Due date to be changed to
unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Creig's status continues to be
uncertain. There will be no way to
react to agreement on naming
conventions before September, hence
this item will be left with an
unspecified due date.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
Status: This matter was discussed at
the Colorado Springs meeting (March,
1995). No consensus has emerged so
the need for the discussion that the
RFP would elicit is probably still
needed.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Status: A wording for this RFP was
proposed by Rich Wrenn at the last
(May, 1995) meeting. This was re-
presented and a couple of
modifications proposed. This item
will be taken over by the PVR group
(Chair: Dave Skinner), and Rich will
submit the RFP text to the reflector
on behalf of the PVR group by the end
of July.
6. Status of the "draft" if the
Reference Model (must it be renewed,
or is it a "guide" and does not need
to be renewed).
Responsible: Merritt Jones.
Due: 95/7 meeting.
Status: Merritt has submitted a
request to the Technical Committee
asking that the Reference Model
document be recognized as a "guide",
and that it be exempted (permanently)
from the expiry clause. This has been
passed on to the IEEE for voting at
the next scheduled meeting. Status of
this request will be reviewed at the
95/9 meeting
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
Status: No change.
5. Status of recruitment and sponsorship
=====================================
As anticipated at last meeting, Dave Skinner presented a
paper at Network Storage 95 (a conference organized by
the other Mike Peterson) brought about half-a-dozen
inquiries to Dave Skinner and Rich Wrenn about the
function of the SSSWG and the role of the Mass Storage
Symposium. Mr. Peterson's organization collects industry
intelligence, focussing mostly on the "industry sweet
spot", rather than high end.
A need was expressed for brochure's etc., for the MSS at
Monterey in order to elevate level and interest and
increase exposure of the work of the SSSWG. No action
taken.
6. Status of issued and open RFP's
===============================
OID
Issued: Bob Baird (VSS)
It was presented at the last meeting that there was one
response and some subsequent discussion. Bob Baird has
updated the response and planned a final report to be put
on the ftp site and/or WEB page. It is not known whether
this has been done.
Base Types
Issued: Mike Peterson (PVL)
Two responses: To be added to the agenda for the next
meeting.
Criteria: consistency requirements between model
components.
Issued: Mike Peterson (PVL)
7. Posix status
============
Dave Skinner met with "Posix.1k" last week. They are
interested in an interface standard to augment or extend
the existing SCSI "robotics" interface. Dave suggests
that this is an appropriate piece of work for P1244. A
new PAR would be required, and would require additional
people interested in this level of work.
Hamid Vazire placed this work in context with what is
currently being done in P1244: The Posix group is trying
to standardize the "mtio" interface for serial access
media. Functional entry points are being defined to take
over operations that are now done through (largely non-
standardized) IOCTL functions. This work does not impact
P1244.
There was discussion of whether there was existing
overlap between the "media changer" interface and PVR
functions. There were views of both sides of this
question, but it was agreed a "clean" layering could be
defined.
There was then a rather complicated discussion about the
layering of OS specific interfaces (such as the Posix
interface) with 1244 interfaces. Bruce and Rich finally
agreed that layering could be done in both directions.
Bruce was volunteered to provide a set of pictures and a
summary of this discussion.
Dave will carry the message to Posix.1k that we support
the work on "mtio.h" and the "mt" utility. He will
continue to recommend that the media changer work should
be done in 1244, but it is not expected that this will
lead to any action at this time.
8. Discussion of "session" abstraction.
====================================
Dave re-presented the content of the paper that he had
earlier distributed via the reflector.
Rich pointed out that "allocate" means to "assign for
exclusive use". "Deallocate" is the release of this
condition. "Reservations" are intended to be essentially
notifications of coming allocations.
Carl described the "ALLOCATE" operation proposed for the
PVL, which is essentially an atomic operation on a set of
resources. It was pointed out that this is essentially a
"reserve and allocate" in one operation.
There was then a discussion of optimization and over-
commitment algorithms, and the need to accommodate legacy
systems, all of which argued in favor of the two step
model. The PVL group said that they would consider these
matters.
Rich was volunteered to summarize the discussion by
preparing material that would address:
1. Updating Dave's summary of the
"session" abstraction to have a
separate "reserve" and "allocate"
operation (rather than the collective
"commit" operation that causes an
"allocate" on every current
reservation.
2. Similarly, an update to the diagram
in Dave's email to represent the
change defined in (1).
3. To draft a statement that defines the
applicability of this updated session
model to each of the components of
the 1244 model
The session adjourned at 12:01 PM, the general session to
be resumed Wednesday morning.
General Session, Wednesday, 19 July, 9.02 am.
9. Discussion of "non-cooperating clients and security".
=====================================================
Bruce K. Haddon re-presented the argument that he had
earlier distributed in his working paper, entitled "Non-
co-operating clients, security, and related issues". The
essential idea was that the requirement for non-co-
operating clients was real and on-going, and that a
"security" methodology could provide a means of achieving
that requirement.
In the course of discussion, it became evident that the
term "non-co-operating" was unfortunate. This term is
used in operating system theory in the context of "co-
operating sequential processes", which relates to the use
of various mechanisms for synchronizing access to shared
resources, and, as such, is a component of the resource
allocation and deadlock avoidance scenario. In the
context of the OSSI model, the term refers simply to
clients that are using the same services, but are not in
communication with each other concerning those services.
The term "mutually unaware" was offered as a better way
to describe this scenario.
It also became apparent that it is the "access control"
aspects of security that are the pieces needed to effect
the creation of an environment in which clients could use
a service and be mutually unaware. So the working paper
presented by Bruce may have been better titled "Mutually
unaware clients, access control, and related issues".
A final agreement was not reached on the issues raised by
Bruce in the working paper. The final position is
summarized as follows:
1. There shall be a model developed for
access control that can be adopted by
each of the 1244 components.
2. Access control is recognized as being
sufficient to satisfy what was termed
in the working paper as the PVL
Requirement, viz.: "The PVL must be
able to guarantee deadlock free
mounts (of cartridges) to the PVL's
clients".
3. There still exists disagreement as to
whether the PVR should be permitted
to have mutually unaware clients.
The session adjourned at 11:44 am, the general session to
be resumed Thursday morning.
General Session, Thursday, 20 July, 8.59 am.
The discussion of "base types" scheduled at this time was
postponed to a time to be determined. Bruce was again
volunteered to update his earlier submission to the
reflector to incorporate more recent thinking.
The session adjourned at 9.15 am, the general session to
be resumed Thursday afternoon.
10. PVL subcommittee report(Mike Peterson)
======================================
Mike reported that the PVL subcommittee had met on the
Monday preceding this meeting, or order to make
additional progress on their draft API, and now, during
the meeting week, had been and were continuing to spend
the major proportion of its time on writing down the
various items upon which the committee had agreed.
He felt that the PVL API had solved some major problems,
but in terms of writing effort, still had a long way to
go, and that it was the PVL subcommittee's intention to
continue working on this at the end of the general
session and on into tomorrow
11. PVR subcommittee report(Dave Skinner)
=====================================
Dave reported that the draft API was in an advanced state
of writing, had been transformed into the desired "man"
page format, and was now being subjected to some overall
organizational corrections.
The main piece to add, which had been the subject of
current discussion, was concerned with security, and the
security interfaces. A proposal would be made for text
covering this subject that would be general in nature, as
there were still issues to be resolved with interactions
with the PVL.
There could be a chance, if it were possible for the PVR
subcommittee to get together between now and September,
for a draft that could be handed out at the MSS.
12. MVR subcommittee report(Bruce K. Haddon)
========================================
Bruce reported on behalf of Creig Humes. A draft of the
MVR API was available, also in the desired "man" page
format, with introductory sections concerning the mover's
objects, and their characteristics. In many ways, the MVR
draft was in the best shape of the three.
However, the question of security, asynchronous events,
and inter-MVR protocols were yet to be addressed.
13. Tabled items
============
None.
14. Review of open action items
======================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Rich Wrenn for Dave
Skinner (MVR).
Due: 7/31.
6. Update of status request for "guide
status" of the draft of the Reference
Model.
Responsible: Merritt Jones.
Due: 95/9 meeting.
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
8. Diagrams showing multiple layering of
MVR's and device drivers and MVR's,
etc.
Responsible: Bruce K. Haddon
Due: unspecified.
9. Summary of issues and refinement of
"session" concept.
Responsible: Rich Wrenn.
Due: unspecified.
10. Re-issue of data representation
response by Bruce.
Responsible: Bruce K. Haddon
Due: 95.9 meeting.
15. Dates
=====
* Dates for 1995
May 16-19: Valley Forge, Pennsylvania. (Host:
Rich Garrison)
July 18-21: DEC, Seattle. (Host: Michael
Peterson)
Sept. 14-15 (Monterey, at the end of the MSS)
Nov. 7-10 (ACSC, Los Angeles, Anit Gafni)
* Dates for 1996 (partial list)
Jan 16-19, Tucson, Arizona (Host: Eric
Stouffler, IBM).
Mar 12-15 (tentative, Colorado)
June (MSS is in Annecy, France). (Note added
later: the 1996 MSS planned for
Annecy in June has been postponed. A
new date is not yet available.)
16. Proposed agenda for 95/9 meeting
================================
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of Action Items from previous meeting
Status of recruitment and sponsorship.
Status of RFP's.
Subcommittee Sessions
Closing General Session
Subcommittee status (each Chairman)
Tabled Items
Review of Action Items
Agenda for 95/11 meeting.
## Resolved: that this proposed agenda be
accepted.
Moved: from the Chair.
Passed unanimously.
17. Other business
==============
None.
Session was adjourned 4 p.m., Thursday, 20 May, 1995.
Received: from wharf.arl.mil by ADMII.ARL.MIL id aa24415; 22 Nov 95 13:12 EST
Received: from stortek.com by wharf.arl.mil id aa12764; 22 Nov 95 13:05 EST
Received: from garonne.stortek.com by stortek.com with SMTP id AA24570
(5.65c/IDA-1.4.4 for ); Wed, 22 Nov 1995 10:21:34 -0700
Received: from gauss.stortek.com by garonne.stortek.com (SMI-8.6/SMI-SVR4)
id SAA28611; Wed, 22 Nov 1995 18:21:17 +0100
Received: by gauss.stortek.com (5.x/SMI-SVR4)
id AA00537; Wed, 22 Nov 1995 18:21:16 +0100
Date: Wed, 22 Nov 1995 18:21:16 +0100
From: "Bruce K. Haddon (Toulouse"
MMDF-Warning: Parse error in original version of preceding line at wharf.arl.mil
Message-Id: <9511221721.AA00537@gauss.stortek.com>
To: cole@arl.mil
Subject: Re: July Meeting Minutes
X-Sun-Charset: US-ASCII
IEEE Storage System Standards Working Group Meeting
18-21st July, 1995
Digital Equipment Corporation, Seattle /
Bellevue, Washington.
Attendance:
Carl Braganza (Advanced Computing Systems
Company, Inc.)
Tim Boyce (IBM, Austin)
Jack Cole (Army Research Laboratory)
Anat Gafni (Advanced Computing Systems
Company, Inc.)
Richard Golding (Hewlitt-Packard Laboratories,
Palo Alto)
Bruce K. Haddon (Storage Technology European
Operations)
Merritt Jones (MITRE Corporation)
Michael T. Peterson (Digital Equipment
Corporation, West)
Dave Skinner (Fujitsu)
Eric Stouffler (IBM Storage Systems Division,
Software Products)
Hamid Vazire (Hewlitt-Packard, Cupertino)
Rich Wrenn (Digital Equipment Corporation,
Co. Springs)
Joel Williams (MITRE Corporation)
Chair: Dave Skinner.
Recording Secretary: Bruce K. Haddon.
General Session, Tuesday 18th July, 9.00 a.m.
Agenda
======
(This agenda reflects minor changes made in the course of
the meeting.)
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of action items from previous meeting
Status of recruitment and sponsorship.
Status of issued RFP's
Posix Update
Discussion of "session" abstraction (Dave E.
Skinner)
Subcommittee Sessions
Discussion of "non-cooperating clients and
security" (Bruce K. Haddon)
Subcommittee Sessions
Discussion of "base types" (Mile Peterson)
Closing General Session
PVL subcommittee report(Mike Peterson)
PVR subcommittee report(Dave Skinner)
MVR subcommittee report(Bruce K. Haddon)
Tabled items
Review of open action items
Dates
Agenda for 95/9 meeting
1. Announcements
=============
"Early registration" for MSS (closes mid-August, please
attend)
Local Seattle meeting details.
Copies available of:
PVR Draft
MVR Draft
"The Role of Standards in Storage Solutions"
presented at Network Storage 95, by Dave
Skinner.
2. Minutes of previous meeting
===========================
It was decided at the May meeting that minutes are to
carry organizational affiliations (and should have this
last time). This will be implemented in the minutes of
the current meeting.
## Resolved: that the minutes of the previous
meeting be approved.
Moved: Mike Peterson; 2nd: Joel Williams.
Carried without dissent.
3. Treasurer's report
==================
Copies of the balance sheet were handed out and
presented. Copies of the presented balance sheets are
available from Dave Skinner. Jack Cole will receive dues
for this meeting.
4. Review of action items from previous meeting
============================================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
Status: No action has been taken. Due
date is to be left unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: 5/31.
Status: Time does not permit
considering this RFP at this time.
Due date to be changed to
unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Creig's status continues to be
uncertain. There will be no way to
react to agreement on naming
conventions before September, hence
this item will be left with an
unspecified due date.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
Status: This matter was discussed at
the Colorado Springs meeting (March,
1995). No consensus has emerged so
the need for the discussion that the
RFP would elicit is probably still
needed.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Status: A wording for this RFP was
proposed by Rich Wrenn at the last
(May, 1995) meeting. This was re-
presented and a couple of
modifications proposed. This item
will be taken over by the PVR group
(Chair: Dave Skinner), and Rich will
submit the RFP text to the reflector
on behalf of the PVR group by the end
of July.
6. Status of the "draft" if the
Reference Model (must it be renewed,
or is it a "guide" and does not need
to be renewed).
Responsible: Merritt Jones.
Due: 95/7 meeting.
Status: Merritt has submitted a
request to the Technical Committee
asking that the Reference Model
document be recognized as a "guide",
and that it be exempted (permanently)
from the expiry clause. This has been
passed on to the IEEE for voting at
the next scheduled meeting. Status of
this request will be reviewed at the
95/9 meeting
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
Status: No change.
5. Status of recruitment and sponsorship
=====================================
As anticipated at last meeting, Dave Skinner presented a
paper at Network Storage 95 (a conference organized by
the other Mike Peterson) brought about half-a-dozen
inquiries to Dave Skinner and Rich Wrenn about the
function of the SSSWG and the role of the Mass Storage
Symposium. Mr. Peterson's organization collects industry
intelligence, focussing mostly on the "industry sweet
spot", rather than high end.
A need was expressed for brochure's etc., for the MSS at
Monterey in order to elevate level and interest and
increase exposure of the work of the SSSWG. No action
taken.
6. Status of issued and open RFP's
===============================
OID
Issued: Bob Baird (VSS)
It was presented at the last meeting that there was one
response and some subsequent discussion. Bob Baird has
updated the response and planned a final report to be put
on the ftp site and/or WEB page. It is not known whether
this has been done.
Base Types
Issued: Mike Peterson (PVL)
Two responses: To be added to the agenda for the next
meeting.
Criteria: consistency requirements between model
components.
Issued: Mike Peterson (PVL)
7. Posix status
============
Dave Skinner met with "Posix.1k" last week. They are
interested in an interface standard to augment or extend
the existing SCSI "robotics" interface. Dave suggests
that this is an appropriate piece of work for P1244. A
new PAR would be required, and would require additional
people interested in this level of work.
Hamid Vazire placed this work in context with what is
currently being done in P1244: The Posix group is trying
to standardize the "mtio" interface for serial access
media. Functional entry points are being defined to take
over operations that are now done through (largely non-
standardized) IOCTL functions. This work does not impact
P1244.
There was discussion of whether there was existing
overlap between the "media changer" interface and PVR
functions. There were views of both sides of this
question, but it was agreed a "clean" layering could be
defined.
There was then a rather complicated discussion about the
layering of OS specific interfaces (such as the Posix
interface) with 1244 interfaces. Bruce and Rich finally
agreed that layering could be done in both directions.
Bruce was volunteered to provide a set of pictures and a
summary of this discussion.
Dave will carry the message to Posix.1k that we support
the work on "mtio.h" and the "mt" utility. He will
continue to recommend that the media changer work should
be done in 1244, but it is not expected that this will
lead to any action at this time.
8. Discussion of "session" abstraction.
====================================
Dave re-presented the content of the paper that he had
earlier distributed via the reflector.
Rich pointed out that "allocate" means to "assign for
exclusive use". "Deallocate" is the release of this
condition. "Reservations" are intended to be essentially
notifications of coming allocations.
Carl described the "ALLOCATE" operation proposed for the
PVL, which is essentially an atomic operation on a set of
resources. It was pointed out that this is essentially a
"reserve and allocate" in one operation.
There was then a discussion of optimization and over-
commitment algorithms, and the need to accommodate legacy
systems, all of which argued in favor of the two step
model. The PVL group said that they would consider these
matters.
Rich was volunteered to summarize the discussion by
preparing material that would address:
1. Updating Dave's summary of the
"session" abstraction to have a
separate "reserve" and "allocate"
operation (rather than the collective
"commit" operation that causes an
"allocate" on every current
reservation.
2. Similarly, an update to the diagram
in Dave's email to represent the
change defined in (1).
3. To draft a statement that defines the
applicability of this updated session
model to each of the components of
the 1244 model
The session adjourned at 12:01 PM, the general session to
be resumed Wednesday morning.
General Session, Wednesday, 19 July, 9.02 am.
9. Discussion of "non-cooperating clients and security".
=====================================================
Bruce K. Haddon re-presented the argument that he had
earlier distributed in his working paper, entitled "Non-
co-operating clients, security, and related issues". The
essential idea was that the requirement for non-co-
operating clients was real and on-going, and that a
"security" methodology could provide a means of achieving
that requirement.
In the course of discussion, it became evident that the
term "non-co-operating" was unfortunate. This term is
used in operating system theory in the context of "co-
operating sequential processes", which relates to the use
of various mechanisms for synchronizing access to shared
resources, and, as such, is a component of the resource
allocation and deadlock avoidance scenario. In the
context of the OSSI model, the term refers simply to
clients that are using the same services, but are not in
communication with each other concerning those services.
The term "mutually unaware" was offered as a better way
to describe this scenario.
It also became apparent that it is the "access control"
aspects of security that are the pieces needed to effect
the creation of an environment in which clients could use
a service and be mutually unaware. So the working paper
presented by Bruce may have been better titled "Mutually
unaware clients, access control, and related issues".
A final agreement was not reached on the issues raised by
Bruce in the working paper. The final position is
summarized as follows:
1. There shall be a model developed for
access control that can be adopted by
each of the 1244 components.
2. Access control is recognized as being
sufficient to satisfy what was termed
in the working paper as the PVL
Requirement, viz.: "The PVL must be
able to guarantee deadlock free
mounts (of cartridges) to the PVL's
clients".
3. There still exists disagreement as to
whether the PVR should be permitted
to have mutually unaware clients.
The session adjourned at 11:44 am, the general session to
be resumed Thursday morning.
General Session, Thursday, 20 July, 8.59 am.
The discussion of "base types" scheduled at this time was
postponed to a time to be determined. Bruce was again
volunteered to update his earlier submission to the
reflector to incorporate more recent thinking.
The session adjourned at 9.15 am, the general session to
be resumed Thursday afternoon.
10. PVL subcommittee report(Mike Peterson)
======================================
Mike reported that the PVL subcommittee had met on the
Monday preceding this meeting, or order to make
additional progress on their draft API, and now, during
the meeting week, had been and were continuing to spend
the major proportion of its time on writing down the
various items upon which the committee had agreed.
He felt that the PVL API had solved some major problems,
but in terms of writing effort, still had a long way to
go, and that it was the PVL subcommittee's intention to
continue working on this at the end of the general
session and on into tomorrow
11. PVR subcommittee report(Dave Skinner)
=====================================
Dave reported that the draft API was in an advanced state
of writing, had been transformed into the desired "man"
page format, and was now being subjected to some overall
organizational corrections.
The main piece to add, which had been the subject of
current discussion, was concerned with security, and the
security interfaces. A proposal would be made for text
covering this subject that would be general in nature, as
there were still issues to be resolved with interactions
with the PVL.
There could be a chance, if it were possible for the PVR
subcommittee to get together between now and September,
for a draft that could be handed out at the MSS.
12. MVR subcommittee report(Bruce K. Haddon)
========================================
Bruce reported on behalf of Creig Humes. A draft of the
MVR API was available, also in the desired "man" page
format, with introductory sections concerning the mover's
objects, and their characteristics. In many ways, the MVR
draft was in the best shape of the three.
However, the question of security, asynchronous events,
and inter-MVR protocols were yet to be addressed.
13. Tabled items
============
None.
14. Review of open action items
======================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Rich Wrenn for Dave
Skinner (MVR).
Due: 7/31.
6. Update of status request for "guide
status" of the draft of the Reference
Model.
Responsible: Merritt Jones.
Due: 95/9 meeting.
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
8. Diagrams showing multiple layering of
MVR's and device drivers and MVR's,
etc.
Responsible: Bruce K. Haddon
Due: unspecified.
9. Summary of issues and refinement of
"session" concept.
Responsible: Rich Wrenn.
Due: unspecified.
10. Re-issue of data representation
response by Bruce.
Responsible: Bruce K. Haddon
Due: 95.9 meeting.
15. Dates
=====
* Dates for 1995
May 16-19: Valley Forge, Pennsylvania. (Host:
Rich Garrison)
July 18-21: DEC, Seattle. (Host: Michael
Peterson)
Sept. 14-15 (Monterey, at the end of the MSS)
Nov. 7-10 (ACSC, Los Angeles, Anit Gafni)
* Dates for 1996 (partial list)
Jan 16-19, Tucson, Arizona (Host: Eric
Stouffler, IBM).
Mar 12-15 (tentative, Colorado)
June (MSS is in Annecy, France). (Note added
later: the 1996 MSS planned for
Annecy in June has been postponed. A
new date is not yet available.)
16. Proposed agenda for 95/9 meeting
================================
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of Action Items from previous meeting
Status of recruitment and sponsorship.
Status of RFP's.
Subcommittee Sessions
Closing General Session
Subcommittee status (each Chairman)
Tabled Items
Review of Action Items
Agenda for 95/11 meeting.
## Resolved: that this proposed agenda be
accepted.
Moved: from the Chair.
Passed unanimously.
17. Other business
==============
None.
Session was adjourned 4 p.m., Thursday, 20 May, 1995.
IEEE Storage System Standards Working Group Meeting
18-21st July, 1995
Digital Equipment Corporation, Seattle /
Bellevue, Washington.
Attendance:
Carl Braganza (Advanced Computing Systems
Company, Inc.)
Tim Boyce (IBM, Austin)
Jack Cole (Army Research Laboratory)
Anat Gafni (Advanced Computing Systems
Company, Inc.)
Richard Golding (Hewlitt-Packard Laboratories,
Palo Alto)
Bruce K. Haddon (Storage Technology European
Operations)
Merritt Jones (MITRE Corporation)
Michael T. Peterson (Digital Equipment
Corporation, West)
Dave Skinner (Fujitsu)
Eric Stouffler (IBM Storage Systems Division,
Software Products)
Hamid Vazire (Hewlitt-Packard, Cupertino)
Rich Wrenn (Digital Equipment Corporation,
Co. Springs)
Joel Williams (MITRE Corporation)
Chair: Dave Skinner.
Recording Secretary: Bruce K. Haddon.
General Session, Tuesday 18th July, 9.00 a.m.
Agenda
======
(This agenda reflects minor changes made in the course of
the meeting.)
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of action items from previous meeting
Status of recruitment and sponsorship.
Status of issued RFP's
Posix Update
Discussion of "session" abstraction (Dave E.
Skinner)
Subcommittee Sessions
Discussion of "non-cooperating clients and
security" (Bruce K. Haddon)
Subcommittee Sessions
Discussion of "base types" (Mile Peterson)
Closing General Session
PVL subcommittee report(Mike Peterson)
PVR subcommittee report(Dave Skinner)
MVR subcommittee report(Bruce K. Haddon)
Tabled items
Review of open action items
Dates
Agenda for 95/9 meeting
1. Announcements
=============
"Early registration" for MSS (closes mid-August, please
attend)
Local Seattle meeting details.
Copies available of:
PVR Draft
MVR Draft
"The Role of Standards in Storage Solutions"
presented at Network Storage 95, by Dave
Skinner.
2. Minutes of previous meeting
===========================
It was decided at the May meeting that minutes are to
carry organizational affiliations (and should have this
last time). This will be implemented in the minutes of
the current meeting.
## Resolved: that the minutes of the previous
meeting be approved.
Moved: Mike Peterson; 2nd: Joel Williams.
Carried without dissent.
3. Treasurer's report
==================
Copies of the balance sheet were handed out and
presented. Copies of the presented balance sheets are
available from Dave Skinner. Jack Cole will receive dues
for this meeting.
4. Review of action items from previous meeting
============================================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
Status: No action has been taken. Due
date is to be left unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: 5/31.
Status: Time does not permit
considering this RFP at this time.
Due date to be changed to
unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Creig's status continues to be
uncertain. There will be no way to
react to agreement on naming
conventions before September, hence
this item will be left with an
unspecified due date.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
Status: This matter was discussed at
the Colorado Springs meeting (March,
1995). No consensus has emerged so
the need for the discussion that the
RFP would elicit is probably still
needed.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Creig Humes (MVR).
Due: Unspecified.
Status: A wording for this RFP was
proposed by Rich Wrenn at the last
(May, 1995) meeting. This was re-
presented and a couple of
modifications proposed. This item
will be taken over by the PVR group
(Chair: Dave Skinner), and Rich will
submit the RFP text to the reflector
on behalf of the PVR group by the end
of July.
6. Status of the "draft" if the
Reference Model (must it be renewed,
or is it a "guide" and does not need
to be renewed).
Responsible: Merritt Jones.
Due: 95/7 meeting.
Status: Merritt has submitted a
request to the Technical Committee
asking that the Reference Model
document be recognized as a "guide",
and that it be exempted (permanently)
from the expiry clause. This has been
passed on to the IEEE for voting at
the next scheduled meeting. Status of
this request will be reviewed at the
95/9 meeting
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
Status: No change.
5. Status of recruitment and sponsorship
=====================================
As anticipated at last meeting, Dave Skinner presented a
paper at Network Storage 95 (a conference organized by
the other Mike Peterson) brought about half-a-dozen
inquiries to Dave Skinner and Rich Wrenn about the
function of the SSSWG and the role of the Mass Storage
Symposium. Mr. Peterson's organization collects industry
intelligence, focussing mostly on the "industry sweet
spot", rather than high end.
A need was expressed for brochure's etc., for the MSS at
Monterey in order to elevate level and interest and
increase exposure of the work of the SSSWG. No action
taken.
6. Status of issued and open RFP's
===============================
OID
Issued: Bob Baird (VSS)
It was presented at the last meeting that there was one
response and some subsequent discussion. Bob Baird has
updated the response and planned a final report to be put
on the ftp site and/or WEB page. It is not known whether
this has been done.
Base Types
Issued: Mike Peterson (PVL)
Two responses: To be added to the agenda for the next
meeting.
Criteria: consistency requirements between model
components.
Issued: Mike Peterson (PVL)
7. Posix status
============
Dave Skinner met with "Posix.1k" last week. They are
interested in an interface standard to augment or extend
the existing SCSI "robotics" interface. Dave suggests
that this is an appropriate piece of work for P1244. A
new PAR would be required, and would require additional
people interested in this level of work.
Hamid Vazire placed this work in context with what is
currently being done in P1244: The Posix group is trying
to standardize the "mtio" interface for serial access
media. Functional entry points are being defined to take
over operations that are now done through (largely non-
standardized) IOCTL functions. This work does not impact
P1244.
There was discussion of whether there was existing
overlap between the "media changer" interface and PVR
functions. There were views of both sides of this
question, but it was agreed a "clean" layering could be
defined.
There was then a rather complicated discussion about the
layering of OS specific interfaces (such as the Posix
interface) with 1244 interfaces. Bruce and Rich finally
agreed that layering could be done in both directions.
Bruce was volunteered to provide a set of pictures and a
summary of this discussion.
Dave will carry the message to Posix.1k that we support
the work on "mtio.h" and the "mt" utility. He will
continue to recommend that the media changer work should
be done in 1244, but it is not expected that this will
lead to any action at this time.
8. Discussion of "session" abstraction.
====================================
Dave re-presented the content of the paper that he had
earlier distributed via the reflector.
Rich pointed out that "allocate" means to "assign for
exclusive use". "Deallocate" is the release of this
condition. "Reservations" are intended to be essentially
notifications of coming allocations.
Carl described the "ALLOCATE" operation proposed for the
PVL, which is essentially an atomic operation on a set of
resources. It was pointed out that this is essentially a
"reserve and allocate" in one operation.
There was then a discussion of optimization and over-
commitment algorithms, and the need to accommodate legacy
systems, all of which argued in favor of the two step
model. The PVL group said that they would consider these
matters.
Rich was volunteered to summarize the discussion by
preparing material that would address:
1. Updating Dave's summary of the
"session" abstraction to have a
separate "reserve" and "allocate"
operation (rather than the collective
"commit" operation that causes an
"allocate" on every current
reservation.
2. Similarly, an update to the diagram
in Dave's email to represent the
change defined in (1).
3. To draft a statement that defines the
applicability of this updated session
model to each of the components of
the 1244 model
The session adjourned at 12:01 PM, the general session to
be resumed Wednesday morning.
General Session, Wednesday, 19 July, 9.02 am.
9. Discussion of "non-cooperating clients and security".
=====================================================
Bruce K. Haddon re-presented the argument that he had
earlier distributed in his working paper, entitled "Non-
co-operating clients, security, and related issues". The
essential idea was that the requirement for non-co-
operating clients was real and on-going, and that a
"security" methodology could provide a means of achieving
that requirement.
In the course of discussion, it became evident that the
term "non-co-operating" was unfortunate. This term is
used in operating system theory in the context of "co-
operating sequential processes", which relates to the use
of various mechanisms for synchronizing access to shared
resources, and, as such, is a component of the resource
allocation and deadlock avoidance scenario. In the
context of the OSSI model, the term refers simply to
clients that are using the same services, but are not in
communication with each other concerning those services.
The term "mutually unaware" was offered as a better way
to describe this scenario.
It also became apparent that it is the "access control"
aspects of security that are the pieces needed to effect
the creation of an environment in which clients could use
a service and be mutually unaware. So the working paper
presented by Bruce may have been better titled "Mutually
unaware clients, access control, and related issues".
A final agreement was not reached on the issues raised by
Bruce in the working paper. The final position is
summarized as follows:
1. There shall be a model developed for
access control that can be adopted by
each of the 1244 components.
2. Access control is recognized as being
sufficient to satisfy what was termed
in the working paper as the PVL
Requirement, viz.: "The PVL must be
able to guarantee deadlock free
mounts (of cartridges) to the PVL's
clients".
3. There still exists disagreement as to
whether the PVR should be permitted
to have mutually unaware clients.
The session adjourned at 11:44 am, the general session to
be resumed Thursday morning.
General Session, Thursday, 20 July, 8.59 am.
The discussion of "base types" scheduled at this time was
postponed to a time to be determined. Bruce was again
volunteered to update his earlier submission to the
reflector to incorporate more recent thinking.
The session adjourned at 9.15 am, the general session to
be resumed Thursday afternoon.
10. PVL subcommittee report(Mike Peterson)
======================================
Mike reported that the PVL subcommittee had met on the
Monday preceding this meeting, or order to make
additional progress on their draft API, and now, during
the meeting week, had been and were continuing to spend
the major proportion of its time on writing down the
various items upon which the committee had agreed.
He felt that the PVL API had solved some major problems,
but in terms of writing effort, still had a long way to
go, and that it was the PVL subcommittee's intention to
continue working on this at the end of the general
session and on into tomorrow
11. PVR subcommittee report(Dave Skinner)
=====================================
Dave reported that the draft API was in an advanced state
of writing, had been transformed into the desired "man"
page format, and was now being subjected to some overall
organizational corrections.
The main piece to add, which had been the subject of
current discussion, was concerned with security, and the
security interfaces. A proposal would be made for text
covering this subject that would be general in nature, as
there were still issues to be resolved with interactions
with the PVL.
There could be a chance, if it were possible for the PVR
subcommittee to get together between now and September,
for a draft that could be handed out at the MSS.
12. MVR subcommittee report(Bruce K. Haddon)
========================================
Bruce reported on behalf of Creig Humes. A draft of the
MVR API was available, also in the desired "man" page
format, with introductory sections concerning the mover's
objects, and their characteristics. In many ways, the MVR
draft was in the best shape of the three.
However, the question of security, asynchronous events,
and inter-MVR protocols were yet to be addressed.
13. Tabled items
============
None.
14. Review of open action items
======================
1. First draft of a process to address
questions of lifetime, size, logging,
evaluation, etc., for RFP's.
Responsible: Dave Skinner and Jack
Cole.
Due: Unspecified.
2. RFP to be issued: "Plug and Play".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
3. RFP to be issued: "Naming
Conventions".
Responsible: Creig Humes (MVR).
Due: Unspecified.
4. RFP to be issued: "API calling
conventions (argument passing)".
Responsible: Dave Skinner (PVR).
Due: Unspecified.
5. RFP to be issued: "Exception Handling
and Error Propagation".
Responsible: Rich Wrenn for Dave
Skinner (MVR).
Due: 7/31.
6. Update of status request for "guide
status" of the draft of the Reference
Model.
Responsible: Merritt Jones.
Due: 95/9 meeting.
7. Copy of the strategic plan outline
and subcommittee strategic plan to be
posted to the reflector.
Responsible: Dave Skinner.
Due: Unspecified.
8. Diagrams showing multiple layering of
MVR's and device drivers and MVR's,
etc.
Responsible: Bruce K. Haddon
Due: unspecified.
9. Summary of issues and refinement of
"session" concept.
Responsible: Rich Wrenn.
Due: unspecified.
10. Re-issue of data representation
response by Bruce.
Responsible: Bruce K. Haddon
Due: 95.9 meeting.
15. Dates
=====
* Dates for 1995
May 16-19: Valley Forge, Pennsylvania. (Host:
Rich Garrison)
July 18-21: DEC, Seattle. (Host: Michael
Peterson)
Sept. 14-15 (Monterey, at the end of the MSS)
Nov. 7-10 (ACSC, Los Angeles, Anit Gafni)
* Dates for 1996 (partial list)
Jan 16-19, Tucson, Arizona (Host: Eric
Stouffler, IBM).
Mar 12-15 (tentative, Colorado)
June (MSS is in Annecy, France). (Note added
later: the 1996 MSS planned for
Annecy in June has been postponed. A
new date is not yet available.)
16. Proposed agenda for 95/9 meeting
================================
General Session
Announcements
Minutes of the previous meeting
Treasurer's report
Review of Action Items from previous meeting
Status of recruitment and sponsorship.
Status of RFP's.
Subcommittee Sessions
Closing General Session
Subcommittee status (each Chairman)
Tabled Items
Review of Action Items
Agenda for 95/11 meeting.
## Resolved: that this proposed agenda be
accepted.
Moved: from the Chair.
Passed unanimously.
17. Other business
==============
None.
Session was adjourned 4 p.m., Thursday, 20 May, 1995.
SSSWG Home