Hi Steve, Thanks for your comments. Although not complete, here are some comments: - We understand the File upload stuff is meant to be a transient/interim and a real repository is expected. - We have one use case with commercial data center for ACS, where clients exist outside of the data center and submit their task to the system, then the task is deployed and executed in the system. In this case, it might require much and/or indeterministic latency for component to pull the data outside of the system. Thus, we assumed that it is more reasonable to have data repository inside the system, even though it may not be what the CDDLM specification requires. We assumed that the purpose of having AddFile is to "push" the data from outside to inside of the system in advance. Then components can "pull" the data easier from there. - As such, the file server in the case 1 diagram may not be a primary or sole implementation of CDDLM, but we understand it is among possible implementations of larger system under the current CDDLM specification. (We may have had to color it neutral one rather than light blue.) - I understood all topics are about the case 1 diagram. Do you have any issues on case 2 or 3? BTW, Thanks for a heads up for JSR 277. We will keep an eye for it. It sounds like the improvement of the jar itself and pretty much java focus. -Keisuke Steve Loughran wrote:
Keisuke Fukui wrote:
Folks in the cddlm-wg,
As you know we will have a joint session at GGF14 between CDDLM and ACS. This is in preparation for the session from acs-wg. After the GGF13 inside the acs-wg, we studied and discussed about the possible interactions between CDDLM and ACS, especially in terms of "File upload" section and AddFile() in the deployment API document. Sequence diagrams in the attachment describe our interpretation about how CDDLM works, and our proposal for the possible interactions in the case that the ACS co-exists in the system. We believe our proposal goes along with what the current set of CDDLM specifications define and doesn't require change in the original definitions.
We are looking forward to discuss about this at the joint session for CDDLM/ACS at GGF14. It is very much appreciated if we get responses before the joint session at GGF14 in case that important overlook in our understanding is found. Please feel free to make comments or questions.
FYI, At GGF14 joint session, we'd like to start our discussion with this diagram and if we don't find critical issues, we may go down to the detail of the interface and/or advanced interaction. We'd also like to hear requirements on ACS in terms of the event notification and asynchronous invocation of the interfaces, if the time permitting.
Thanks in advance for you efforts on this!
Best Regards, Keisuke Fukui ACS-WG
thank you, I will comment briefly.
-The File upload stuff was very much written to be a transient/interim solution in the absence of a real repository, which is why it is so minimal and barely functional. I didnt want to be dependent upon anything not yet designed, but didn't want to do a repository myself
-the current revision allows the sender to declare the URL schema to use. That could be file: for a shared filesystem and http: or https for HTTP access. It could also be fancy custom stuff; there is some java project whose name escapes me that provides a multicast URL resolution system, the file could be stored across multiple machines without ever having to give them a hostname, which is very good for fault-tolerance.
-There is also support for adding metadata to a request, but nothing to do searches, retrieval, or even introspection on what stuff is supported. I've left that for implementations.
-There is the perennial problem of how to get files up over SOAP. SwA is only available in java distrubutions, and then not consistently, DIME is in .NET WSE and Axis 1,x, but even more unpopular. As for MTOM, well, who implements that yet?
As a workaround I've put in stuff for having the endpoint actually retrieve the files themselves, but this is a bit unsatisfactory, because it requires the files to be broadly visible on the 'net, and introduces race conditions. Otherwise, data goes inline in base-64 encoded form, unless/until MTOM lets you pretend that the file attached to the message is really inline base 64.
here are two example mesages and responses from the XSD-validation tests
<api:addFileRequest> <api:name>urn://45</api:name> <api:mimetype>application/x-pdf</api:mimetype> <api:schema>file</api:schema> <api:uri>http://example.org/files/source.pdf</api:uri> </api:addFileRequest> </t:test>
<api:addFileResponse> <item>file://nas1/temp/source.pdf</item> <item>file://nas2/4fgdbb.tmp</item> </api:addFileResponse>
<api:addFileRequest> <api:name>urn://46</api:name> <api:mimetype>application/x-ssh-key</api:mimetype> <api:schema>http</api:schema> <api:data> AAAAB3NzaC1yc2EAAAABIwAAAIEAwVmUkPzXdWEyJZ8nCR8GvdrDtO00RI4Z Bg3Gyviuz5IrWj2C6b2BdcKn+S/swDV1fiEFY4+ewYHUfmg+UKm2T8Lfksjn Hinks0GoVvkwy3bF48U5yVk1akAzR5YbSLJa6Naj8XS9681xVzWpbjxrV3KR QNWvEqI0MqRE34MzT4M= </api:data> <api:metadata> <x:expires date="2005-07-18" xmlns:x="http://example.org/expiry" /> </api:metadata> </api:addFileRequest> <api:addFileResponse> <item>http://server/job5/files/1</item> </api:addFileResponse>
On a related note, I see that Fujitsu are listed as one of the interested parties in JSR 277, the java modules/repository proposal. Are you involved in that?
-steve