DAIS Telcon - 03/05/05 ====================== Chair : Norman Paton Notes : Mario Antonioletti Attendees: ---------- Mario Antonioletti, EPCC Dave Pearson, Oracle Norman Paton, University of Manchester Allen Luniewski, IBM Susan Malaika, IBM Hemant Kolwalkar, IBM Simon Laws, IBM Agenda: ------- Review Relational Schema Model for CIM. Actions: -------- [Norman] Create a list of options for update of disconnected data sets. [Simon] Write up the top level IO operation to go in the core spec. [Mario/Simon/Amy] Look at DAIS issue 1200 and either decide what needs to done to resolve this and respond accordingly on Grid Forge or flag it as an issue that needs to be discussed at some future telcon. [Mario] Continue working on the DAIS contribution to the data architecture. [Susan] Look into how the XML attributes and elements for the resource properties relate to the DMTF CIM model. ----- Key points for discussion: - (1) Determine what has to be done so that the CIM model and resulting schema are acceptable to DAIS. - (2) Walk through the current model. - Note: do not have XML schemas yet - no consensus regarding the style or tooling to be used to generate the XML schema. Summary Of Follow Up Items Needed: - Go after a namespace for the XML representation of the properties of the resources in the model - Clarify namespace scope (entry point, version etc) - Give feedback to DMTF on XML style required for DAIS - Clarify how vendor specific extensions work in the CIM model - Clarify where capabilities go See discussion from the call below ................................................. Default speaker is Susan Malaika. Can we just rely on a name space or do we require a schema? Norman: will not block the DAIS progress regarding the generation of the XML schema but it would be nice to have some input into its generation. Likely to be a while before the XML schema is sorted. Trying to get agreement on a namespace. Can give feedback/input into the XML Schema through a white paper. Dave: do we think that the document that Susan and Norman wrote gives enough direction for the DMTF to specify schema? Not that one. We do not want XMI style XML. The schemas will be fairly flat, need to determine how they are wrapped - the root element - and how we would use them. There is a second document which I've worked on - not public yet - which could have some examples. Some extra pieces that we would like to the CIM model. There is a change request underway at the DMTF to try to define where the relational mdel ties into the current CIM model ... *A rational representation of the model is presented and a walk through is made* Norman: to what extent do vendors follow an SQL 99 standard in their data dictionary? The CIM model can include vendor specific extensions. Dave: from what is being described vendors will be able to map a vanilla SQL99 on to this model. Can vendors extend the model though in a way which reflect specific vendor implementations? That is "business as usual" for the DMTF. Dave: but there needs to be a standard way of generating extensions though so that users can interpret the data in the extensions. The UML and the MOF that the DMTF produces allows for vendor specific extensions. Dave: I'm sure that DMTF provides standard mechanisms but they have to try and define them in the same way. I think that there is a typing mechanism - there is some form of extensions. Will make sure that this is very clear. ... Norman: the more of the SQL 99 spec that gets incorporated the better. Also the case that it would be fine for DAIS if this came through in steps. Completeness is good but a rolling program would be alright but the sooner this is done the better. Do we expect one XML schema for the whole thing? and people would issue XPath into the whole thing? Where does the relevant part of the CIM model start? We could make a recommendation for DAIS to start off at the SQL object root. Norman: that would be fine. Hemant: just cleaning up the model at the moment Mario: is the namespace going to be used as a versioning mechanism? That is another decision that we need to make - would have to specify what version of the CIM model is used. Dave: that is something that we will need to identify - what version of the CIM model is being used. Do we want to tie the DAIS spec to the SQLObject and then a version? Dave: we need an attribute that gives us the version. Norman: do not want to tie down to a given version. We need to determine what version is being used then... Norman: we need to understand what versioning and naming scheme they will use. Dave: what about the resource capabilities? Does the current definition include those descriptions? No, seen as being outside the SQL99 model. Dave: DMTF have started to define this ... capabilities are required then Dave: with distributed queries it is important to be able to determine this. are we saying that we would like the DMTF to take this on-board? Dave: we ought to make them aware that we are interested in information that describes the capabilities and that we would also be willing to give input about what capabilities are important. Do we have a place where capabilities could go? Hemant: there is something in the catalogue... We would have to see where capabilities can go, if there is a good place already. Norman: worth noting, something needs to be said in the DAIS specs about extensions to the CIM model - if capabilities were to be included in the CIM model then one could provide this information in DAIS ... it might be possible to have hooks where we put metadata which state what bits of the CIM model this DAIS implementation populates The whole model being in one namespace? Norman: just how people populate ... the version of the product that you run behind the scenes. You might - or might not - want to expose that. So SQL Object may not be the root as you might be interested in another things? Norman: we are interested in several distinct things. This has some bearing on how you approach the whole thing. Dave: I agree, it would be interesting to see if CIM offers different entry points that you might want to investigate. So, we just want an XML representation without any management stuff. Simon: we are just interested in the types. Will circulate my document in a few days and update the model and make it public on the DAIS web site. Will give checkpoints on progress. Mario: so do we pull the explicit CIM stuff from the relational spec. Norman: yes.