Re: [kmip] Groups - kmip-spec-v1.1-csprd01-attribute-index-update-summary.pdf uploaded

[Tim Hudson <tjh@cryptsoft.com>]

On 6/04/2012 3:53 PM, John Leiseboer wrote:

All,

 

Comments below on attribute index, as requested at TC meeting.

The Attribute structure defined in 2.1.1 is used in two entirely separate contexts:
1) returning the value of an attribute
2) specifying an attribute as a parameter to an operation

In context 2 there are three sub-contexts:
2a) a specific known attribute index must be specified
2b) the attribute index is not able to be specified as it has not yet been created
2c) the attribute index shall not be specified as all instances or any instance of the attribute are referenced

Add Attribute is 2b
Modify Attribute and Delete Attribute is 2a
Get Attributes and Locate is 2c (although Locate actually specifies that if an attribute index is provided it SHALL be ignored).

In the context of 1 and 2a where an attribute index is allowed to be specified then if a value is not specified then it is assumed to be 0.
In the context of 2b and 2c an attribute index value is not allowed to be specified.

As I said on the TC call the objective in the changes I was making to revert the introduction of attribute index of -1 from KMIP 1.1 was focused on simply addressing those additions without expanding out to change areas the text which existed in KMIP 1.0 - i.e. I was limiting the scope of the change as much as possible.

I'll expand the proposal to include sorting out the single instance / multi-instance issues that we have previously discussed now that it is clear the acceptable scope of change is broader than the parameters we discussed on previous calls.

There are many areas within KMIP which I think we all agree could have been handled differently; but a minor specification update is not the right context in which we should be making fundamental incompatible changes to core concepts.

Tim.