randys@otl.SanDiego.NCR.COM (Randy Smerik) (11/21/89)
I have a few questions on LU 6.2: 1) The LU 6.2 TPRM describes the SYNCPT and BACKOUT verbs. I have heard that these two verbs were never really implemented except internally within CICS. Is this true? 2) Is there any vendor that implemented the SYNCPT and BACKOUT verbs? If so, what platforms, etc. 3) Is it true that the SYNCPT/BACKOUT verbs were not carried forward to SAA CPIC? 4) I have heard the phrase "IBM 2-phase commit protocols" used. How do they relate to SYNCPT/BACKOUT? What are they? Thanks in advance, Randy Smerik, randy.smerik@sandiego.ncr.com, (619) 485-2084
dukew@binky.sybase.com (Wayne Duquaine) (11/22/89)
(1) SYNCPT and BACKOUT were also implemented on the S/38 I believe, in
addition to CICS.
(2) Any other vendor implemented SYNCPT and BACKOUT ? - I think
CULLINET SOFTWARE's UCF tp monitor may have also implemented these
as part of it LU 6.2 support. It runs on S/370 mainframes.
(3) SYNCPT and BACKOUT were not carried forward in SAA CPIC ? - sad
but true, and I regard this as a major screwup by IBM. I heard
part of the problem was due to inter-product fights while CPIC
was being designed. I've heard rumors that an officially blessed
version of SAA CPIC SYNCPT and BACKOUT will appear in some
new version of the CPIC manual in the next 1-2 years.
(4) How do 2-phase commit protocols relate to SYNCPT and BACKOUT ? -
Application (transaction) programs issue SYNCPT and BACKOUT calls.
These calls are then translated into the required (SNA) 2-phase
commit protocol (prepare/commit/committed/forget) over the wire.
For more information on how these work, you may want to read
Chapter 5.3 "Presentation Services - Sync Point Services Verbs"
in IBM's SC31-6808 LU 6.2 Reference: Peer Protocols manual.
Hope this helps.
Wayne D.
=== Usual disclaimers apply ===dukew@binky.UUCP (Wayne Duquaine) (11/22/89)
(1) SYNCPT and BACKOUT were also implemented on the S/38 I believe, in
addition to CICS.
(2) Any other vendor implemented SYNCPT and BACKOUT ? - I think
CULLINET SOFTWARE's UCF tp monitor may have also implemented these
as part of it LU 6.2 support. It runs on S/370 mainframes.
(3) SYNCPT and BACKOUT were not carried forward in SAA CPIC ? - sad
but true, and I regard this as a major screwup by IBM. I heard
part of the problem was due to inter-product fights while CPIC
was being designed. I've heard rumors that an officially blessed
version of SAA CPIC SYNCPT and BACKOUT will appear in some
new version of the CPIC manual in the next 1-2 years.
(4) How do 2-phase commit protocols relate to SYNCPT and BACKOUT ? -
Application (transaction) programs issue SYNCPT and BACKOUT calls.
These calls are then translated into the required (SNA) 2-phase
commit protocol (prepare/commit/committed/forget) over the wire.
For more information on how these work, you may want to read
Chapter 5.3 "Presentation Services - Sync Point Services Verbs"
in IBM's SC31-6808 LU 6.2 Reference: Peer Protocols manual.
Hope this helps.
Wayne D.
=== Usual disclaimers apply ===wclark@csi.3Com.Com (Wayne Clark) (11/28/89)
In article <2107@ncr-sd.SanDiego.NCR.COM> randys@otl.SanDiego.NCR.COM (Randy Smerik) writes: > >I have a few questions on LU 6.2: > >1) The LU 6.2 TPRM describes the SYNCPT and BACKOUT verbs. I have > heard that these two verbs were never really implemented except > internally within CICS. Is this true? Support for Synpoint Services (aka Option Set 108 per the latest update to the TPRM) was announced on November 1st for VTAM 3.3. It will be available in June, 1990. I have never encountered any other implementations of Syncpoint Services outside CICS. > >2) Is there any vendor that implemented the SYNCPT and BACKOUT verbs? > If so, what platforms, etc. Not sure. I do know that none of the major vendors of SNA OEM software have Syncpoint Services. > >3) Is it true that the SYNCPT/BACKOUT verbs were not carried forward > to SAA CPIC? True. Stay tuned. > >4) I have heard the phrase "IBM 2-phase commit protocols" used. How > do they relate to SYNCPT/BACKOUT? What are they? The term "2-phase" is used because there are two distinct phases that the synchronized Transaction Programs must go through to complete the distributed transaction: one to verify that all TP's in the distributed transaction are ready to commit and one to actually make the changes happen. > >Thanks in advance, >Randy Smerik, randy.smerik@sandiego.ncr.com, (619) 485-2084 One final note: the OSI Distributed Transaction Processsing (DTP) model currently under standardization looks very similar to LU 6.2. (Not merely coincidental since IBM played a major role in its definition.) However, as might be expected, the OSI DTP model contains some features not found in LU 6.2, especially in the area of distributed transaction processing. I don't believe you will see much activity from IBM in the area of LU 6.2 Syncpoint Services until after OSI DTP reaches International Standard level, which is supposed to be in the Autumn of 1990. Then the LU 6.2 Syncpoint Services model may be altered to match OSI DTP. ------------ Wayne Clark IBM Connectivity Operations, 3Com Corporation (formerly Communications Solutions, Inc.) Phone: 408/562-6967 EMail: Wayne_Clark@SPD.3mail.3Com.COM -or- nsc,bnrmtv,epimass}!csi!wclark
wclark@csi.UUCP (Wayne Clark) (11/28/89)
In article <2107@ncr-sd.SanDiego.NCR.COM> randys@otl.SanDiego.NCR.COM (Randy Smerik) writes: > >I have a few questions on LU 6.2: > >1) The LU 6.2 TPRM describes the SYNCPT and BACKOUT verbs. I have > heard that these two verbs were never really implemented except > internally within CICS. Is this true? Support for Synpoint Services (aka Option Set 108 per the latest update to the TPRM) was announced on November 1st for VTAM 3.3. It will be available in June, 1990. I have never encountered any other implementations of Syncpoint Services outside CICS. > >2) Is there any vendor that implemented the SYNCPT and BACKOUT verbs? > If so, what platforms, etc. Not sure. I do know that none of the major vendors of SNA OEM software have Syncpoint Services. > >3) Is it true that the SYNCPT/BACKOUT verbs were not carried forward > to SAA CPIC? True. Stay tuned. > >4) I have heard the phrase "IBM 2-phase commit protocols" used. How > do they relate to SYNCPT/BACKOUT? What are they? The term "2-phase" is used because there are two distinct phases that the synchronized Transaction Programs must go through to complete the distributed transaction: one to verify that all TP's in the distributed transaction are ready to commit and one to actually make the changes happen. > >Thanks in advance, >Randy Smerik, randy.smerik@sandiego.ncr.com, (619) 485-2084 One final note: the OSI Distributed Transaction Processsing (DTP) model currently under standardization looks very similar to LU 6.2. (Not merely coincidental since IBM played a major role in its definition.) However, as might be expected, the OSI DTP model contains some features not found in LU 6.2, especially in the area of distributed transaction processing. I don't believe you will see much activity from IBM in the area of LU 6.2 Syncpoint Services until after OSI DTP reaches International Standard level, which is supposed to be in the Autumn of 1990. Then the LU 6.2 Syncpoint Services model may be altered to match OSI DTP. ------------ Wayne Clark IBM Connectivity Operations, 3Com Corporation (formerly Communications Solutions, Inc.) Phone: 408/562-6967 EMail: Wayne_Clark@SPD.3mail.3Com.COM -or- nsc,bnrmtv,epimass}!csi!wclark