MPLS, GMPLS and Pseudowires: The Need for Cooperation Between Standards Bodies Adrian Farrel Old Dog Consulting [e-mail]

1
MPLS, GMPLS and PseudowiresThe Need for
Cooperation Between Standards Bodies Adrian
FarrelOld Dog Consultingadrian_at_olddog.co.uk
www.mpls2005.com
2
Why Have Standards?

• Do we all want standardization?
• Sometimes I wonder!
• Good for Providers
• More choice and flexibility
• Good for Vendors
• Ease of market penetration
• Good for Individuals
• Transferable skill set
• All of these points have equally powerful
  opposites

3
Why Have Standards Bodies (SDOs)?

• What is the alternative?
• Each company develops and documents its own
  solution
• Everyone has to implement two or three solutions
• Chaos, confusion and expense
• We can learn from other people
• There is a relatively small global pool of
  experts
• Acapulco is very nice at this time of year
• Does everyone in the SDO want standardization?

4
Which SDOs Impact MPLS and GMPLS?

• Internet Engineering Task Force (IETF)
• International Telecommunications Union (ITU)
• Optical Interworking Forum (OIF)
• Institute of Electrical and Electronic Engineers
  (IEEE)
• MPLS, ATM and Frame Relay Alliance (MFA)
• Tele Management Forum (TMF)
• IP Sphere ??
• Others ??

5
Some Overlap Between SDOs

• This is an understatement!
• Converged networks imply converged SDOs
• Examples
• IETF, ITU-T and OIF working on protocol
  specifications for control planes for optical
  networks
• IETF and ITU-T overlap on MPLS applications such
  as
• OAM
• Pseudowires
• MPLS transport
• IETF examining ways to manage IEEE Ethernet
  networks
• MFA and ITU-T working on MPLS NNIs etc.

6
Are Multiple MPLS and GMPLS SDOs Helpful?

• Against
• Do we want multiple standards for the same
  function?
• Can we hope to attend all relevant standards
  meetings?
• SDOs become marketing organizations
• Chaos, confusion and expense
• For
• Different SDOs have developed specializations
• Different applications, protocols or functions
• Architecture versus protocol
• Requirements, experimentation, development,
  operation
• Different SDO processes are conducive to
  different results
• Its good for job creation
• Hawaii is a pretty nice place

7
What Is the Solution?

• SDOs have a duty of care to the industry
• No SDO should regard itself as better than any
  other
• Each SDO must protect itself from abuse
• Standards shopping
• Corporate politics
• Each SDO must be aware of overlaps and potential
  overlaps
• Communicate plans
• Seek natural homes for work
• Cooperate to a common goal

8
What Form Should Cooperation Take?

• Focus on the end result
• Devolve work to areas of expertise
• Clear sense of ownership of protocols
• Understanding of wider implications of changing
  protocols
• Responsiveness between SDOs
• Questions and clarifications
• Requests for assistance or work
• Formal liaison process may be helpful

9
Do We Need Formal Processes?

• Skepticism, lack of trust, and turf wars
• Bad history
• Over-specified process may ensure smooth
  operation
• Can relax the process in the future
• Process must include a commitment to follow the
  process
• For example
• MPLS and GMPLS Change Process (draft-andersson-rt
  g-gmpls-change-01)
• Makes the IETF the place for developing protocol
  extensions for MPLS and GMPLS
• Includes a commitment by the IETF to receive
  requirements from other SDOs