Header - request or reply
       - version #

Subject - src/dest/single (for, say, finding the loss on a particular interface)
           - either - ip addresses
                    - MP names

Methodology - "tool" (e.g. delay, delay.twoway, delay.twoway.owamp) - maybe ask for lots of tools
                 - ip settings (ip version, size, tos, traffic pattern) }  ,- take the default
                 - transport settings (tcp/udp, port)                   } o                ,- mandatory
                 - application (use H323 pattern, etc)                  }  `- give a value -- optional
                 - time (start,end)                                     }
                 - statistics (raw, average...)                         }
                    - parameters (time intervals, amount of data...)    }

===

Phone conf

Mark, Eric, Jeff, Nicola, Lokic, Me

Eric:
-----
We use this call as an extension of our meetings at GGF plus at Internet2.
Want to define the "problem", if not a first-draft solution.


Nicola:
-------
The NMWG document can only contain 1 result per NetworkMeasurement.
So they are leaning towards being only able to request from one path section at a time.

Eric: agrees that you should ask for a link at a time rather than all the parts of a whole path.

Describes the stuff we discussed.



Dante's first draft request doc can be used for making a measurement or just requesting data.

There are two responses: either you get an ID to refer to the measurement being made; or you get the results.

Eric: We need to be able to write some of this stuff up in text, so that we can say what abilities we require.
eg: "we would like the schema to ask the following questions, in our proposed schema this is how you would be able to do them".
Then we have the parameters of what we need to achieve.

Need to write up a list of use cases.