WG RAQMON Internet-Drafts

1
WG RAQMON Internet-Drafts

• RMON MIB WG Meeting
• Washington, Nov. 11, 2004

2
New Round of IETF Drafts as IETF-60 discussions

• TCP transport
• Draft-ietf-rmonmib-raqmon-framework-07.txt
• New metrics, existing metrics changes,
  clarifications
• Draft-ietf-rmonmib-raqmon-pdu-07.txt
• Add TCP transport
• Fixes, changes, clarifications as per framework
  changes
• Draft-ietf-rmonmib-raqmon-mib-05.txt
• Fixes, changes, clarifications as per framework
  changes
• Boilerplate changes for new Intellectual Property
  RFCs

3
Draft-ietf-rmonmib-raqmon-framework-07.txt

• Framework and main RAQMON entities definition
• Requirements
• For RDS
• RRC
• Transport Protocol
• RAQMON Data Model
• Metrics changes, as per IETF-60 discussions
• Rename / redefine / define new metrics for Round
  Trip End-to-End Network Delay, One Way End-to-End
  Network Delay, Application Delay, Inter-Arrival
  Jitter, IP Packet Delay Variation, Total Number
  of Application Packets Received, Total Number of
  Application Packets Sent, Total number of
  Application Octets Received, Total number of
  Application Octets Sent, Cumulative Application
  Packet Discards, Packet discards in Fraction,

4
RAQMON Data Model

• Roundtrip End-to-End Network Delay
• One way End-to-End Network Delay
• Application Delay
• Inter Arrival Jitter
• IP Packet Delay Variation
• Source Payload Type
• Receiver Payload Type
• Total number of Packets Received
• Total number of Packets Sent
• Total number of Octets Received
• Total number of Octets Sent
• Cumulative Packet Loss
• Cumulative Packet Discard
• Packet Loss in Fraction (in )
• Packet Discard in Fraction (in )

• Data Source Name (DN)
• Receiver Name (RN)
• Data Source Address (DA)
• Receiver Address (RA)
• Data Source Device Port used
• Receiver Device Port used
• Application Name

1
4

• Session Setup Date/Time
• Session Setup delay
• Session duration
• Session Setup Status

2

• Source Layer 2 Priority
• Destination Layer 2 Priority
• Source Layer 3 Priority
• Destination Layer 3 Priority

3

• CPU utilization in Fraction (in )
• Memory utilization in Fraction (in )

5
add other parameters by using extension of
Vendor Specific part of PDU
5
Comments and Open issues wrt. draft-ietf-rmonmib-r
aqmon-framework-07.txt

• 5.7 Session Setup Date/ Time
• Should be titled "Report Date/ Time" as it
  relates to the time at which the report was
  generated rather than the session setup
• Resolve text inconsistency
• Should this be in time zone independent format to
  permit easier correlation on large networks?
• No, NTP format as per RFC 1305 is widely deployed
  operationally
• 5.8 Session Setup Delay
• "The Session Setup Delay metric reports the time
  taken from an origination request being initiated
  by an endpoint to the media path being
  established (or a call progress indication being
  received from the remote endpoint.)
• accept

• 5.11/ 5.12 End-to-End delay
• Add a note to say that the packets used for
  measurement may be of a different type to those
  used for media (e.g. ICMP instead of RTP) and
  hence may differ in terms of route and queueing
  priority. This may result in measured delays
  being different to those experienced on the media
  path.
• Accept work on clarification text
• 5.12 - last paragraph
• it would be simpler and more logical to say that
  RAQMON implementations should NOT derive one way
  delay by dividing rtd by 2 - just leave the
  parameter out if it is not known.
• Accept

6
Comments and Open issues wrt. draft-ietf-rmonmib-r
aqmon-framework-07.txt (2)

• 5.13 Application delay
• "The network delay metrics described in sections
  5.11 and 5.12"......"The Application Delay
  metrics defined in this section are intended to
  capture additional elements of delay"
• it is not clear if it is intended that the
  application delay includes BOTH encoding and
  buffer/decode delay, or are there two parameters?
• Clarification receiving end delays
• it is also confusing to talk about this as
  application delay, which should really be
  end-end. Call it "application endpoint delay",
  or add network delay and endpoint delay and call
  the sum "application delay
• See above

• 5.16 etc
• Some IP endpoints separate signaling and media
  path system components. It would be more
  practical to say that applications packets MAY
  include signaling packets
• Accepted
• 5.20 Cumulative Packet Loss
• Since there is now a packet discard count then it
  is easier to state that a late packet may be
  classed as discarded or lost - there should be no
  ambiguity
• Clarify avoid double counting
• Mandatory use of SNMP
• Clarified on list with the commenter
• RDS may use one of the transport
• RRC MUST support both

7
Draft-ietf-rmonmib-raqmon-pdu-07.txt

• Combined tcpsip I-D with the RAQMON PDU draft
• Draft renamed to reflect multiple transport
• Has two options supported in PDU Draft
• Native TCP
• SNMP
• RDS can implement either one, but RRC MUST
  implement both
• Syntactical re-arrangement of PDU to accommodate
  4 New Parameters

• End to End Delay Split
• Application/End Device Delay
• Network Delay (RTT OWD)
• Cumulative Packet Discards
• Discards (in )
• Corrections to the MIB as per metrics changes
• change template for new IPR requirements
• IANA considerations

8
PDU Structure
Shortened Report Type
Re-Arrangement 256 Sub sessions
..
..
Extensions
9
Draft-ietf-rmonmib-raqmon-mib-04.txt

• change syntax of objects in raqmonParticipantTable
  to Integer32 to add values of -1 in cases when
  the collector did not receive any report on the
  specific metrics
• change object names to align with
  RAQMON-FRAMEWORK
• added objects in raqmonParticpantTable to cover
  all metrics in RAQMON-FRAMEWORK
• added raqmonRDSTimeout object to control the
  timeout for reports in collectors
• change template for new IPR requirements
• aligned REFERENCE clauses with new numbering in
  RAQMON-FRAMEWORK
• added new mandatory IANA considerations section

10
Conclusions and Recommendations

• WGLC comments editorial in nature
• Comments resolution includes clarifications in
  the framework, no impact on PDU and MIB documents
• We seem to be converging on content and quality
• Its been taking so long (11th IETF meeting!)
• Recommend to forward to the IESG, with the edits
  as per decisions of this meeting