WEBI Issues List : 2002-04-01

From: WEBI Issues List (webi@equinix.com)
Date: Mon Apr 01 2002 - 11:18:45 MST


WEBI Issues List
Last update: '$Date: 2002/03/18 22:04:59 $'
-------------------------------------------------------------------
1 - Cache coherence examples
      Created: 2002-02-20
        Locus: RUP-REQS
        Topic: General
     Priority: Editorial
       Status: New
Originator(s):
               - Ian Cooper <mailto:ian@the-coopers.org>
     Owner(s):
  Description:
      Introduction includes references to cache invalidation
      or cache coherence protocols (WCIP, PSI, DOCP).

--
      Should we give examples at all? If so are there other
      references that need to be given?

------------------------------------------------------------------- 2 - CONSISTENCY GUARANTEE has no definition Created: 2002-02-25 Locus: RUP-REQS Topic: Term Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: The term CONSISTENCY GUARANTEE is used within the body of the document but there is no description of this term.

------------------------------------------------------------------- 3 - RESOURCE GROUP has no definition Created: 2002-02-25 Locus: RUP-REQS Topic: Term Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: The term RESOURCE GROUP is used within the body of the document but there is no description of this term.

------------------------------------------------------------------- 4 - Design Guidelines necessary? Created: 2002-02-25 Locus: RUP-REQS Topic: General Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Is the Design Guidelines section necessary? Most of these guidelines should be common sense to designers of candidate protocols (e.g. if it won't scale then it doesn't belong on the Internet).

------------------------------------------------------------------- 5 - RUP CLIENT invalidation examples Created: 2002-02-25 Locus: RUP-REQS Topic: Usecase Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: In the RUP CLIENT driven invalidation text, is it appropriate to give examples upon which the RUP SERVER basis its decisions? E.g. should we mention Etag, timestamp etc.?

------------------------------------------------------------------- 6 - Content location update analagous to HTTP 30x? Created: 2002-02-25 Locus: RUP-REQS Topic: Usecase Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Is a content location update analagous to an HTTP 30x response? -- If so is it useful to indicate this in section Content location update (6.3)?

------------------------------------------------------------------- 7 - Content location update analagous to RFC3143 A.1.1? Created: 2002-02-25 Locus: RUP-REQS Topic: General Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Is content location update analagous to the archived Known Problem presented in section A.1.1 of RFC3143?

------------------------------------------------------------------- 8 - Content location update temporary? Created: 2002-02-26 Locus: RUP-REQS Topic: Requirement Priority: Design Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: How temporary is a content location update?

------------------------------------------------------------------- 9 - Content location update granularity Created: 2002-02-26 Locus: RUP-REQS Topic: Requirement Priority: Design Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: What is the level of granularity of a content location update? -- Current wording suggests a per-entity basis, but if a reasonable encoding could be provided it may be more beneficial to enable a resource group to have its location updated.

------------------------------------------------------------------- 10 - Metadata updates section needed Created: 2002-02-26 Locus: RUP-REQS Topic: Requirement Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: No text for metadata updates. Do we need this section? If so we need text describing what metadata updates are, even if we say they SHOULD NOT be implemented.

------------------------------------------------------------------- 11 - Loose consistency? Created: 2002-02-26 Locus: RUP-REQS Topic: Requirement Priority: Design Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Does this section belong in the requirements and guidelines? -- It's not clear how this RUP metadata would be communicated or how it would interact with CLIENT and SERVER invalidations. It sounds like it's a way of overriding HTTP's Cache-Control directives while allowing consistency of 0 to n second resolution on a group of entities or a resource group.

------------------------------------------------------------------- 12 - HTTP Warning codes? Created: 2002-02-26 Locus: RUP-REQS Topic: General Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Which HTTP warning code should be used to indicate content is being served in a manner HTTP says is stale but which is indicated as fresh due to RUP control? -- (draft-01 mentioned 110/111, but should we register a new warning code?)

------------------------------------------------------------------- 13 - Limit on only announcing updates to subscribed groups unworkable Created: 2002-02-27 Locus: RUP-REQS Topic: Requirement Priority: Design Status: Active Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: (draft-02) 6.2 item 1 stated that RUP SERVERs must only send notifications about resources to RUP CLIENTs that have requested notifications for that group. -- RUP CLIENTs (that have not yet seen entities within a group) cannot, therefore, benefit from content location updates [which could be used in the future] or content prefect hints [the RUP Server may know that there is anticipated demand]. -- Changing MUST to SHOULD (or even including the words in non-RFC2119 form) does not seem to work.

------------------------------------------------------------------- 14 - Option acknowledgement a problem with relays Created: 2002-02-27 Locus: RUP-REQS Topic: Requirement Priority: Design Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Option acknowledgement of operations seems impossible in a situation where a relay is in use unless communication between RUP SERVER and the relay contains additional information, and is actually a different or superset of the protocol.

------------------------------------------------------------------- 15 - Security considerations needs major work Created: 2002-02-27 Locus: RUP-REQS Topic: Security Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: Security considerations is still in note form and the editor needs some help in fleshing this out.

------------------------------------------------------------------- 16 - References Created: 2002-02-27 Locus: RUP-REQS Topic: General Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: References [10] - [17] are never used, and I suspect serve "reading list" purposes as a couple of others in the list do. Do we need these references? Do we need all the other references for this document to make sense?

------------------------------------------------------------------- 17 - Issues list being ignored Created: 2002-03-18 Locus: RUP-REQS Topic: General Priority: Editorial Status: New Originator(s): - Ian Cooper <mailto:ian@the-coopers.org> Owner(s): Description: No-one is commenting on any of the issues.



This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:23:01 MST