[rus-wg] draft-ggf-wsi-rus-14

[Jon MacLaren]

John, Steven,

I read draft v13, and only found stuff about the writing, rather than  
the technical details - good job!  (Oh, and maybe it's just me, but  
the section numbering in v14 seems messed up?)


Security Considerations - most of this is pretty waffly, and I think  
could be replaced with something much shorter.  It also contains a  
couple of downright dangerous statements, such as "The data stored in  
the RUS must also be non-repudiatable".  That's simply not the  
responsibility of the RUS.  If resource managers want to push a bunch  
of incorrect records into the RUS, the implementation is certainly  
not going to stop them...

Also, maybe replace the paragraph summarising the authorisation model  
with the simple sentence "Operations also require authorisation; this  
is discussed in Section 3.1".  Repeating explanations is unnecessary  
and usually leads to inconsistency.

Sec 2.1 (Architecture).  The main purpose of the RUS is stated as  
being to record the consumption of resources during the invocation of  
a Web Service.  But the RUS in the spec is only able to store  
UsageRecords, which only describe resources consumed by computational  
jobs.  I don't think these two reconcile.  And we have operations  
like "extractUsageBySubmitHost"!

How about starting with: "The RUS's purpose is to provide the storage  
and retrieval of resource usage information relating to the execution  
of computational jobs.  Information is transfered in UsageRecords, as  
defined in [UR-SPEC]" or similar.

I think I'd also mention the UsageRecord work in the Abstract.  It's  
fundamental to this work.  (Could just grab the text at the start of  
Section 4.)

Introduction of UsageRecord.  The UsageRecord is mentioned in Section  
3, first paragraph, but not defined until the start of section 4.   
(Should be defined the first time it is mentioned.  See prev. comment  
too.)

XML Schema Fragments  A few of these just appear without being  
prefaced by any explanation.  (Also, some are described as "XML  
fragments" rather than XML schema fragments.)

Sec 5.2 - desc of RUSUserNotAuthorisedFault.  The wording of this is  
a bit ambiguous: "the user has no permission to invoke operations on  
the RUS".  Sounds to me like they don't have permission to do  
anything at all (but surely they might get this when trying to  
insert, and might have permission to extract).  Suggest replacing it  
with: "the user does not have permission to invoke the operation  
they've just attempted" .

Sec 4. Example Usage Record.  A real, valid DN, rather than  
"test_user" would be good...


That's all.  I'll give it another look when it reaches public  
comment, though.

Cheers,

Jon.


On Jul 28, 2005, at 3:24 AM, John Ainsworth wrote:

> New version attached, again I have aligned the schemas and wsdl in  
> the appendix with changes in the main body, and is response to  
> Steven's comments.
>
>> 4. Page 22 changed a RUSId & RecordId to RUSRecordId in the XML  
>> schema -
>> John do you want to update your schema files?
>>
> Yes have done so.
>
>
>> 5. Page 22 should the NextRUSId and HashCode elements be there and
>> should urwg:Usage be urwg:UsageRecord?
>>
>
> NextRUSId and HashCode elements have been removed - they were left  
> over from the Manchester RUS implementation. (Similarly I have  
> removed the operations retrieveConfiguration and updateConfiguration.)
>
> The urwg schema contains the following definitions:
>     <xsd:element abstract="true" name="Usage"  
> type="urwg:UsageRecordType"/>
>     <xsd:element name="UsageRecord" substitutionGroup="urwg:Usage"  
> type="urwg:UsageRecordType"/>
>     <xsd:element name="JobUsageRecord" substitutionGroup="urwg:Usage">
>         <xsd:complexType>
>             <xsd:complexContent>
>                 <xsd:extension base="urwg:UsageRecordType"/>
>             </xsd:complexContent>
>         </xsd:complexType>
>     </xsd:element>
>     <xsd:element name="UsageRecords">
>         <xsd:complexType>
>             <xsd:sequence>
>                 <xsd:element maxOccurs="unbounded" minOccurs="0"  
> ref="urwg:Usage"/>
>             </xsd:sequence>
>         </xsd:complexType>
>     </xsd:element>
>
> John
>
> <draft-ggf-wsi-rus-14.doc>
>
> On 26 Jul 2005, at 12:27, Steven Newhouse wrote:
>
>
>> Right... I think we're getting there now.
>>
>> New draft attached. I've accepted a lot of long standing changes in
>> order to tidy up the text.
>>
>> 1. Added in MUST, MUST NOT, etc. text.
>> 2. Section 5.5.2 - Sven are you happy with the XML?
>> 3. 5.6.1 & 5.6.2 changed RUSResult to OperationResult
>> 4. Page 22 changed a RUSId & RecordId to RUSRecordId in the XML  
>> schema -
>> John do you want to update your schema files?
>> 5. Page 22 should the NextRUSId and HashCode elements be there and
>> should urwg:Usage be urwg:UsageRecord?
>>
>> Any comments from anyone else on the list?
>>
>> Steven
>>
>> -- 
>> ----------------------------------------------------------------
>> Dr Steven Newhouse                        Tel:+44 (0)2380 598789
>> Deputy Director, Open Middleware Infrastructure Institute (OMII)
>> Suite 6005, Faraday Building (B21), Highfield Campus,
>> Southampton University, Highfield, Southampton, SO17 1BJ,  UK
>> <draft-ggf-wsi-rus-13.doc>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/rus-wg/attachments/20050728/a638682e/attachment.htm