Document: draft-irtf-nmrg-snmp-measure-01.txt
Shepherd: Bert Wijnen
Result:	  15 last call reviews, all supportive
Token:    produce new revision

In the following, the phrase "editorial comments" includes suggestions
for improvements that the editor considers non-controversial. A more
detailed writeup will be provided when the document gets revised.

(1) Bert Wijnen:

- Name of the protocol operations (inform -> inform-request,
  trap2->snmpV2-trap)

- Editorial comments

(2) Aiko Pras:

(3) Frank Strauss:

- Likes to see a figure, but it is not yet totally clear what he
  misses

- Questions requirement to keep original pcap data

- Likes to see more information on anonymization, especially for
  merging traces; be clear whether this is in scope or out of scope

- Why decode protocol operations to keywords, and keep error-status as
  a numeric value?

- Editorial comments

(4) Remco van de Meent:

(5) Giorgio Nunzi:

- Storage of original raw traces may be a burden (sensitive data)

- Likes to have an XML schema definition "somewhere" (not necessarily
  in the document - but then where?)

(6) Wes Hardacker:

- Likes to see warnings about interpretation of incomplete/biased data
  sets and encouragement of data providers to document major events
  during the data colletion period

- Several editorial comments

(7) Liam Fallon:

- Editorial comments

(8) Sharon Chisholm:

- Likes to see XML schema

- Clarifications on some trace analysis ideas

- Editorial comments

(9) David Perkins:

- Data type mapping looses information (we use the more minimalistic
  SMIng base type set and not the SMIv2 set of base data types)

- Opaque and netAddr

- Confusion about IPv6 addresses

- Fancy OID formats

- Editorial comments

(10) Deep Medhi:

- Add example for XML and CSV

- Information about typical trace sizes and storage space requirements

- Editorial comments

(11) Randy Bush:

- Questions usefulness of XML format

- Questions the usefulness of section 2.5 (processing traces)

- Questions whether section 3 belongs into the document at all

- Editorial comments

(12) David Harrington:

- Suggests to relax the requirement for long term storage of sensitive
  raw traces

- Suggests several warnings for potential mis-interpretations due to
  biased data sets etc.

- Editorial comments

(13) Dan Romascanu:

- Editorial comments

(14) Luca Deri:

- Why not used (anonymized) pcap files directly?

- Likes to see more text on anonymization

- Editorial comments

(15) Mark Burgess:

- Editorial comments