Sampling%20and%20Filtering%20Techniques%20for%20IP%20Packet%20Selection%20-%20Update%20-%20%20draft-ietf-psamp-sample-tech-04.txt

1
Sampling and Filtering Techniques for IP Packet
Selection- Update - draft-ietf-psamp-sample-tec
h-04.txt

• Tanja Zseby, FhG FOKUS
• Maurizio Molina, NEC Europe Ltd.
• Fredric Raspall, NEC Europe Ltd.
• Nick Duffield, ATT Labs

2
Changes

• Inclusion of 2 mandatory hash functions
• Mandatory if hash-based selection is implemented
• Result of an investigation on various hash
  functions ? Nicks presentation
• Inclusion of extended table for clarification of
  the categorization of schemes
• Text on why configuration of non-uniform
  probabilistic and flow state sampling is not
  detailed
• Operating time removed from the info model
• Small terminology changes (consistent with FW
  draft)
• Some re-wording, minor changes, etc.

3
Hash Functions

• Purpose
• Sampling ID (random sampling emulation)
• Digest ID (packet identifier, e.g. for consistent
  packet selection at multiple observation points)
• Hash Function Definition
• Needed for consistent sampling and identification
• Different requirements for Sampling and Digest ID
• Tests uniformity of distribution, collision
  probability, speed
• Evaluation Result
• IP Shift-XOR (IPSX) for Sampling ID
• CRC32 for Digest ID
• ? MUST for IPSX and CRC32

4
Categorization Table

• Selection Scheme deterministic
  content- Category
• selection
  dependent
• -----------------------------------------
  ------------------
• systematic X
  _ Sampling
• count-based
  
• -----------------------------------------
  ------------------
• systematic X
  - Sampling
• time-based
  
• -----------------------------------------
  ------------------
• random -
  - Sampling
• n-out-of-N
  
• -----------------------------------------
  ------------------
• random -
  - Sampling
• uniform probabilistic
  
• -----------------------------------------
  ------------------
• random -
  (X) Sampling
• non-uniform probabil.
  
• -----------------------------------------
  ------------------
• random -
  (X) Sampling

5
Configuration of non-uniform probabilistic and
flow state sampling

• Many different specific methods possible
• Input parameters dependent on sampling goal and
  implementation
• Some concrete proposals from research
• e.g. EsVa01,DuLT01,Moli03
• But still in early stage
• Need further investigations to prove their
  usefulness and applicability
• ?Specification of explicit schemes left to
  vendors (e.g. as extension of the information
  model)

6
Operating time removed from the info model

• Operating time
• Start/Stop time of sampling process
• Configuration on higher layer (activation of
  scheme)
• Need not be reported (new scheme would be
  announced anyway)
• ? removed
• Further Issues
• Term random sampling approximation remains
• Minor re-wording for clarification

7
Thank you for your attention !

• Questions ?