Re: FSM changes for the Draft-15

[Susan Hares <[e-mail]>]

Alex:

The files I sent use a table format for this approach.   I was planning to
provide an accompanying document with this information in it.
We are just trying to correct the specification, I went with the 2nd 
document idea
as a "compromise"

I'd be glad to go and translate the state table document to your format, but
I'd like some additional feedback whether this will be acceptable at this
stage of the IDR document.

Sue

At 01:56 PM 11/9/2001 -0800, Alex Zinin wrote:

>Sue,
>
> > I'd like 2nd to ask that the working group accepts this FSM rewrite.
>
> > I've re-written the state machine to be a bit cleaner in text.
> > The Idle state had two sub-states so I cleaned this up by making each
> > of these states a different text (Idle and Idle Hold)
>
> > Here are changes to the draft-15 FSM to make the FSM a bit clear.
>
>If we are to make the FSM textual description in the spec more or less
>meaningful, I suggest we take some time and fix the section using
>the following approach:
>
>8. BGP Finite State machine.
>
>8.1 BGP FSM Events
>
>For each event:
>
>  Event       : <event-name>
>  Description : <short-description>
>  Generation  : <conditions when the event is generated>
>
>8.2 BGP Session States
>
>For each state, give a generic description, something like
>the following:
>
>  State       : <state-name>
>  Description : <short description>
>  Valid events: <list of events expected and processed in this state,
>                 others should be ignored>
>  In actions  : <list of actions _always_ performed when
>                 transitioning to the state>
>  Out actions : <list of actions _always_ performed when
>                 transitioning from the state>
>  State actions: <list of actions performed (e.g. periodically) while
>                 in this state>
>
>8.3 BGP FSM Definition
>
>  FSM State : <one or many states>
>  FSM Event : <one or many events>
>  Next state: <new state if always the same>
>  Actions   : <list of state/event-specific actions and state
>               transitions if condition-dependent>
>
>I'll take a look at the files you sent as well.
>
>Alex.