[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: failure detection




El 18/08/2005, a las 14:08, Paul Jakma escribió:

On Thu, 18 Aug 2005, Paul Jakma wrote:

- reachability detection itself is unidirectional

 (iljitsch's draft mentions "unidirectional", i know, but wrt
  outside unidirectional issues. I'm proposing the reachability
  mechanism itself be unidirectional in nature. Two unidirectional
  paths == bi-directional :) )

I'd should explain this better:

- the protocol would be to implement "I can hear you", rather than
  "We can both hear each other" (as, eg, some other path-state
  protocols do - like OSPF Hello)

- As an optimisation, if the other side asks "Can you hear me?", you
  can at least infer "I can hear them" - and hence suppress a probe.

Aside: The address tuples used in each direction could be different


right see Jari's draft for more details

the difficutly is how to deal with unidirectional paths
but imho the mechanism included in Jari's draft can deal with this.

regards, marcelo


PS: I know all the reachability/probing/ULP-feedback talk is all still at exploratory stages. Hope I've not been /too/ presumptive in attacking ULP-feedback and n^2 probing (ie, attacking these when they're only being /considered/, rather than decided on. Just trying to nip these two particular ideas in the bud - at least for anything other than appendix 'implemention note' type material).

regards,
--
Paul Jakma	paul@clubi.ie	paul@jakma.org	Key ID: 64A2FF6A
Fortune:
I can't live the button-down life like you.  I want it all!  The
terrifying lows, the dizzying highs, the creamy middles!

		-- Homer Simpson
		   Lisa's Rival