Re: [MEXT] offline Re: binary filters draft 0

[Michael Eriksson <michael.eriksson at ericsson.com>]

On Wed, May 13, 2009 at 04:55:32 -0700, Giaretta, Gerardo wrote:
> > It is of course an advantage that the same format is used by the
> > connection manager as in the MIP protocol. The connection manager will
> > create an array of bytes (the flow rules in ASCII format). All that
> > the MIP daemon needs to do is to put an MH transport header on this
> > byte array and send it to the HA.
> > 
> > The MIP daemon does not need to do complex binary packing of invidual
> > rules, it will just transfer a byte vector. Trivial to implement!
>
> You are assuming an API which is not standardized so I could claim
> that other systems use different encoding in the internal
> communication.

By using a text format, you have the /possibility/ to have the same
format between the connection manager and the MIP daemon as in the MIP
signaling. This makes simple but powerful implementations possible,
but you are of course free to use any API you want between the
connection manager and the MIP daemon.

I still think that it is elegant and a relevant advantage that the
rules are expressed in a "high level" text language that is easy for
the connection manager to generate, and that the MIP daemon only have
to do very little to transfer the rules to the MN.

> I don't think this should be consider as a differentiating point
> between the solutions as this is clearly not generic enough and only
> based on your personal assumptions.

I can't see how my "personal assumptions" comes into this. I only show
that a text format can nicely tie together a connection manager with
MIP (or any other mobility mechanism).


Regards,
Michael