jwf@hpctdlb.HP.COM (John Freeborg) (08/17/89)
I don't know if I'm reading ISO 8327 correctly or not (the session layer protocol spec.). Here is a question that I'm fuzzy on: Whenever a category 2 SPDU is sent will a category 0 SPDU ALWAYS be concatenated onto the front? (category being used with respect to concatenation categories 0,1, and 2) (the above question being asked irregardless of the extended concatenation option) Basically I'm wondering if the only legal way you'll see a data SPDU is with a give token or please token SPDU concatenated onto the front. With both the give token and data SPDUs having a SI of 1 the spec. would be ambigious if the above rule didn't hold - correct? Can you ever have a TSDU with JUST a data SPDU contained in it? Thanks for any help in advance, John Freeborg ------------------------------------------------------------------------------ John W. Freeborg Hewlett Packard, Corp. jwf@hpcsos.col.hp.com Colorado Telecommunications Division Disclaimer = 5070 Centennial Blvd. 'The opinions expressed are not Colorado Springs, CO 80919 necessarily those of my employer.' (719) 531-4771
vlee@ugly.cs.ubc.ca (viola po-ying lee) (08/18/89)
> > Whenever a category 2 SPDU is sent will a category 0 SPDU > ALWAYS be concatenated onto the front? I read the CCITT 1984 X.225 Recommendation. For concatenation, a category 0 SPDU is always in front of category 2 SPDU(s). As for the valid basic concatenation of DT SPDU, the first pdu is GT SPDU ONLY. PT is not allowed with DT SPDU. There is some rule about this GT. Token item parameter is only present in the GT SPDU if this DT SPDU contains a complete SSDU or the last the segment of a segmented SSDU. (Pages 315-318 of the X.225 document give a good description about valid concatenation) >Can you ever have a TSDU with JUST a data SPDU contained in it? A DT SPDU can be mapped to a TSDU, but it would be considered to be an invalid PDU by the receiver. QUESTION for the ISO Session Layer gurus: Why are some SI values of several SPDUs the same (e.g. MAA and AEA)? If the purpose of SI values is to identify the SPDU type, shouldn't they be all unique? It would be a lot easier to decode a SPDU had the SI values been all unqiue. Viola Lee viola@ubc-idacom.ubc.ca (don't use the vlee@ugly address in the header)
Christian.Huitema@MIRSA.INRIA.FR (Christian Huitema) (08/18/89)
Viola, The rules for SPDU concatenation are better understoud after reading the specs of the Teletex Session (CCITT T.62). A lot of the problems that you mention derive from a CCITT requirement to maintain a compatibility between the single layer ISO session protocol and the two layers (exchange + document) T.62 protocols. Then, there is the point of major synchronisation and activity end: should be considered as two names for the same thing. Christian Huitema
hagens@CS.WISC.EDU (08/18/89)
> > > > Whenever a category 2 SPDU is sent will a category 0 SPDU > > ALWAYS be concatenated onto the front? > I read the CCITT 1984 X.225 Recommendation. For concatenation, a > category 0 SPDU is always in front of category 2 SPDU(s). As for the > valid basic concatenation of DT SPDU, the first pdu is GT SPDU ONLY. > PT is not allowed with DT SPDU. There is some rule about this GT. Token > item parameter is only present in the GT SPDU if this DT SPDU contains > a complete SSDU or the last the segment of a segmented SSDU. (Pages > 315-318 of the X.225 document give a good description about valid > concatenation) No! The GT SPDU is always prepended to a DT SPDU. The token item parameter is present in the GT SPDU iff the ss-user wishes to transfer tokens. If the token item parameter is not present, then the GT has no effect (although it still takes up 2 bytes. > >Can you ever have a TSDU with JUST a data SPDU contained in it? If you mean, just a DT SPDU, the answer is no. You always concatenate. > A DT SPDU can be mapped to a TSDU, but it would be considered to > be an invalid PDU by the receiver. Correct. Don't do this. > QUESTION for the ISO Session Layer gurus: > Why are some SI values of several SPDUs the same (e.g. MAA and AEA)? > If the purpose of SI values is to identify the SPDU type, shouldn't > they be all unique? It would be a lot easier to decode a SPDU had the > SI values been all unqiue. The session protocol was designed to be bit compatable with a pre-exsistent protocol (T.61). This ambiguity of SPDU SI values is a result of this. It is possible to distiniguish the PDUs by looking at the parameters that are present. Rob Hagens
iso@NIC.DDN.MIL (08/20/89)
>Organization: Hewlett-Packard CTD, Colo. Spgs.
I don't know if I'm reading ISO 8327 correctly or not (the session
layer protocol spec.). Here is a question that I'm fuzzy on:
Whenever a category 2 SPDU is sent will a category 0 SPDU
ALWAYS be concatenated onto the front?
(category being used with respect to concatenation categories 0,1, and 2)
(the above question being asked irregardless of the extended concatenation
option)
Basically I'm wondering if the only legal way you'll see a data SPDU is
with a give token or please token SPDU concatenated onto the front.
With both the give token and data SPDUs having a SI of 1 the spec. would
be ambigious if the above rule didn't hold - correct?
Can you ever have a TSDU with JUST a data SPDU contained in it?
Thanks for any help in advance,
John
Freeborg
------------------------------------------------------------------------------
John W. Freeborg Hewlett Packard, Corp.
jwf@hpcsos.col.hp.com Colorado Telecommunications Division
Disclaimer = 5070 Centennial Blvd.
'The opinions expressed are not Colorado Springs, CO 80919
necessarily those of my employer.' (719) 531-4771
viola@idacom.cs.ubc.ca (Viola Lee) (08/22/89)
In article 473, I said: > I read the CCITT 1984 X.225 Recommendation. For concatenation, a > category 0 SPDU is always in front of category 2 SPDU(s). As for the > valid basic concatenation of DT SPDU, the first pdu is GT SPDU ONLY. > PT is not allowed with DT SPDU. There is some rule about this GT. Token > item parameter is only present in the GT SPDU if this DT SPDU contains > a complete SSDU or the last the segment of a segmented SSDU. To which Rob Hagens replied: >>No! >>The GT SPDU is always prepended to a DT SPDU. The token item parameter >>is present in the GT SPDU iff the ss-user wishes to transfer tokens. >>If the token item parameter is not present, then the GT has no effect (although >>it still takes up 2 bytes. Yes! Actually, I quoted almost directly from X.225 (CCITT Red Book, Fascicle VIII.5, page 316, footnote 2): ... the Token Item parameter is only present in the GIVE TOKENS SPDU if this DATA TRANSFER SPDU contains a complete SSDU, or the last segment of a segmented SSDU. I believe ISO 8237 is *very* similar to CCITT X.225, but if it says something different please enlighten me. I would be really interested in widening my view if my information is incorrect. >>It is possible to distiniguish the PDUs by looking at the parameters that >>are present. It is not possible for the MAA and AEA SPDUs. They have exactly the same parameters. Although they generate the same responses at the receiver in almost all states, there are two exceptions: STA04A (AWAIT MAA) and STA04B (AWAIT AEA). Thus, distinguishing by parameters is not always possible. Viola Lee
rick@GATEWAY.MITRE.ORG (08/23/89)
>> It is not possible for the MAA and AEA SPDUs. They have exactly the same >> parameters. Although they generate the same responses at the receiver in >> almost all states, there are two exceptions: STA04A (AWAIT MAA) and STA04B >> (AWAIT AEA). Thus, distinguishing by parameters is not always possible. >> Viola Lee >True, but if you don't implement major sync (since nobody -- last time I >checked) >uses it, then you don't have the problem... >rob Virtual Terminal uses major sync for non-data-transfer functions such as renegotiation of VT parameters during an association. -Rick Wilder