kehres@TIS.LLNL.GOV (Tim Kehres) (11/09/88)
Below is the current text for the gateway workshop report from the recent IFIP conference in Costa Mesa. This report was compiled by Mike Westbrook and myself and we believe the contents to accurately reflect what was discussed in the meeting. This text will be sent to North Holland for publication with the conference proceedings. If there are any errors or omissions, please let me know as soon as possible so that the corrections can be incorporated into the final text. Tim Kehres ---------- cut here ---------- cut here ---------- cut here ---------- MHS Gateway Working Group Report October 11, 1988, Costa Mesa, California, U.S.A. A working group met at the IFIP 6.5 working conference to discuss X.400(84) and X.400(88) interworking problems. Those in attendance were Brad Benson, Jaime Delgado, Tim Kehres, Hanne Larsen, Erik Lillevold, Arne Litlere, Yarran Lu, Jim McHugh, Sonu Mirchandani, Bob Purvy, Michael Rosin, Steve Schramm, Norry Shimizu, Alnoor Shivji, and Mike West- brook. The group used the position paper by Tim Kehres, Message Handling Systems: Interoperability Problems Between 1984 and 1988 X.400 as the basis for most of the discus- sions. The meeting started out by trying to identify the prob- lems with MHS(84) and MHS(88) interworking. The CCITT/ISO recommendations state how this is to be accomplished in Annex B of X.419. They specify reasons for non-delivery and a downgrading function for the MTS Transfer Protocol (P1), but give no recommendations for interworking of the 1984 and 1988 versions of the Interpersonal Messaging Protocol (P2). The group felt that X.419, Annex B does not provide an adequate means for the interworking of 1984 and 1988 sys- tems. Some reasons for this are that the downgrading func- tion for P1 will substantially reduce the quality of service for 1988 users when messages are routed through a 1984 MTA. In addition, there are no guidelines for the interworking of P2(84) and P2(88). In order to solve some of the interworking problems, a gateway type of solution was proposed. The group felt that a non-gateway type of solution would be preferable, but no suitable proposals were presented. At this time, the group split into two sub-groups to investigate the interworking problems associated with the P1 and P2 protocols. The P1 sub-group agreed that the rules defined in Annex B of X.419 for non-delivery were acceptable and should be followed. The discarding of protocol elements in X.419 downgrading was determined not to be acceptable when both end systems were 1988 MHS. When routing a message through a 1984 MTA, this downgrading reduces the quality of service to November 9, 1988 - 2 - the 1988 end systems to that of the 1984 routing MTA, which was determined to be unacceptable. One recommendation is to try to avoid all 1984 MTA's whenever possible, but this clearly is not always practical. The other proposal was to construct a gateway between 1984 and 1988 MTA's which would save the discarded protocol elements in a gateway-specific body part. This gateway would exist in each 1988 system that communicates with a 1984 MTA. The gateway would save the discarded protocol elements when transferring to a 1984 domain, and would re-insert such elements into the P1 envelope when received back into a 1988 domain. The P2 sub-group identified differences between the body part types allowed by P2(84) and P2(88). P2(84) defines the Telex and Simple Formattable Document body part types which are no longer included in P2(88). In addition, P2(88) defines the Bilaterally Defined and the Externally Defined body part types which are not present in P2(84). In the case of a Telex(84) body part arriving in a 1988 MTA, this should be converted to an IA5 body part type using the ITA2 character set. No conversion in the 1988 to 1984 direction is needed. A lot of interest was expressed in the handling of ODA body parts in both 1984 and 1988 systems. The procedures for the conversion of the Simple Formattable Document Type, Bilaterally Defined body part type, and the Externally Defined body part type are for further study. When the two sub-groups met, a recommendation was made that MTA routing algorithms should take downgrading into account, and if possible route around 1984 MTA's, instead of downgrading to route through them. It was also felt that Annex B of X.419 violated the spirit of X.400 since the emphasis was on non-delivery at the domain boundary rather than delivery across domains. A solution needs to be iden- tified that is transparent to the 1984 routing MTA's, but at the same time provides some level of pass-through capability when 1988 end systems are communicating via 1984 relays. This group feels that such a solution would best be met by the specification of a gateway between 1984 and 1988 Message Handling Systems. There were several work items that were identified for further study. They are: o The need for more specific O/R Address downgrading procedures. o The need for further definition of P2 interworking problems and potential solutions. o The need to define a gateway body part type to hold November 9, 1988 - 3 - downgrade-removed service elements. o The need to define rules for the packaging and unpack- aging of discarded P1(88) protocol elements. An internet mailing list has been set up to discuss these issues. Contributions to the list can be submitted by sending messages to: ifip-gtwy@tis.llnl.gov Administrative requests such as subscription requests for addition and deletion from the list can be sent to: ifip-gtwy-request@tis.llnl.gov November 9, 1988
mhsc@oce.nl (11/16/88)
Hello Tim I read your minutes of a IFIP meeting on MHS 84/88 interworking. The use of ODA body part tags was also mentioned in this paper. We are cooperating in a European research project, funded under the Esprit Programme, where we do pilot implementations of ODA. It was not exactly clear to me how your group intends to handle ODA body parts but I would like to explain our position. At the last Hannover Trade Fair in Germany ICL, Bull, Olivetti, Siemens and Oce performed a demonstration of ODIF exchange over an X.400 network. Our experience in this demonstration learned that the simple use of a body part tag is not suffi- cient. There is a need to know the Document Architec- ture and the Document Application Profile used in the document. These matters have also been signaled by the CCITT and in the 1988 revision of the standard there will be a somewhat different handling of ODA body parts. Therefore MHS experts of the 11 companies that consider to participate in the 1989 ODA demonstration have come together. The companies involved include the 5 demons- trators of this year and further Nixdorf, British Telecom, C-labs, Apple Europe and IBM European Network- ing Center. They set up a more detailed scenario for the use of 84 X.400 for ODIF document transfer. These recommendations are based on both the CCITT X.400 and the ISO MOTIS standards and have taken into account a smooth transition towards the 1988 revision of the X.400 standard. Our recommendations have as little as possible effect on the MHS itself. Our approach for conveyance of ODA by MHS IPMS The following is proposed as a medium term solution, i.e. for the lifetime of MHS based on the 1984 recom- mendations of the CCITT. When 1988 based system take over their place, a smooth transition is possible with the solution sketched in this proposal. For P2: An ODA document will be transferred as a single body part with tag 12: oda [12] IMPLICIT OCTET STRING The content of the Octet string will contain (using ASN.1 basic encoding ) a value of type OdaBodyPart, which is a Sequence containing Parameters and Data components: OdaBodyPart ::= SEQUENCE { OdaBodyPartParameters, OdaData } The parameters and Data components are each aligned to Annex E of T.411 (1988) (ie ISO 8613-1) In this way it is up to the ODA implementation to worry about the handling of specific ODA content information. We envisage that to ODA implementation can send a reply message when it is not able to handle the specific Document Application Profile that is carried in the body, just like you would want an 1988 MHS implementation to refuse such a message. I think that besides our configuration there are no ODA/X.400 systems operational for the moment so I don't think that implementations which dont't use OdaBodyPartParameters need much attention. I hope this can be an input to your discussions and I would gladly receive more detailed information of your ideas on the handling of ODA body parts. Maarten Schoonwater PODA project team Oce Nederland Venlo Netherlands mhsc@oce.nl or ...!mcvax!oce-rd1!mhsc