STPP Slide 1

1
STPPSimpleTDMPseudowireProtocol

• PWE3 57th IETF
• 16 July 2003

Yaakov (J) Stein
2
The Great TDM Battle
draft-anavi-tdmoip-00 was prepared for 50th
IETF comparisons presented at 52nd IETF we have
been discussing for gt 2 years several
compromises reached and rejected the
time has come for a decision
50 Minneapolis 51 London 52 Salt Lake City 53
Minneapolis 54 Yokohama 55 Atlanta 56 San
Francisco 57 Vienna
3
draft-anavi-tdmoip-05

• Multiple main modes to optimize for special cases
• raw (STPP)
• AAL1 unstructured, structured, structured w/ CAS
• AAL2
• HDLC
• Various special mechanisms to cope with network
  defects
• adaptive clock recovery
• common clock SRTS
• packet loss avoidance
• packet loss concealment
• Simple interworking with existing AAL1 and AAL2
  networks

4
draft-vainshtein-cesopsn-06

• Evolved into a long complex draft
• Single main mode with plethora of submodes
• framelength locking OR frame locking
• full multiframe in packet OR multiframe marking
• RTP OR RTP compressed OR RTP suppressed
• T1 octet-aligned OR 8-consectuive frames
• optional ECMP Prevention protector
• signaling packets OR in-packet signaling
  structure
• Support (for which submodes?)

5
STPP

• Simple (dumb) TDM pseudowire protocol
• Draft is only 11 pages, readable by non-TDMers
• Designed to cover most common TDM cases
• clock available or unimportant
• low packet loss
• TDM structure unimportant
• CAS/CCS signaling unimportant
• Simply encapsulate arbitrary length TDM packets
• When special sizes are needed, their use is
  encouraged
• default sizes are those needed by CESoPSN
  implementations
• should use sizes that simplify AAL1-interworking
  if needed
• No special payload treatment (e.g. AAL1 buffering
  or frame-locking)
• Highest possible BW efficiency (no extraneous
  overhead)

6
Proposal

• Accept STPP for most common (simplest) TDM cases
• broad support
• easy to implement
• Accept both TDMoIP and CESoPSN for special cases
• structure critical
• CAS signaling
• high packet loss
• problematic clock recovery
• interworking with existing networks
• and let the market decide