Re: [babel] babel Digest, Vol 17, Issue 5

Tony Przygienda <tonysietf@gmail.com> Fri, 06 January 2017 17:50 UTC

MIME-Version: 1.0
In-Reply-To: <878tqogkda.wl-jch@irif.fr>
References: <mailman.1581.1483630892.3886.babel@ietf.org> <CA+wi2hM0AeR684ouWv3Rc8kHnCOEdtDVVLjQT3V4+jQCDHwNZg@mail.gmail.com> <878tqogkda.wl-jch@irif.fr>
From: Tony Przygienda <tonysietf@gmail.com>
Date: Fri, 06 Jan 2017 09:50:05 -0800
Message-ID: <CA+wi2hN9eB3xqQEtUwAHxmrBAJh79h+2c+tBVoEwhFnABjKdjg@mail.gmail.com>
To: Juliusz Chroboczek <jch@irif.fr>
Content-Type: multipart/alternative; boundary="94eb2c1cd62c4e0090054570a8f2"
Archived-At: <https://mailarchive.ietf.org/arch/msg/babel/cw44D1EODctgVzbZ-nNgHwsWlYs>
Cc: Babel at IETF <babel@ietf.org>
Subject: Re: [babel] babel Digest, Vol 17, Issue 5
Precedence: list

See, just needed some prodding ;-)

I agree with everything except maybe the complexity, a fully fledged BGP
these days is 10x a fully fledged IGP. But then again, it's apples and
kiwis, IGPs do not attempt to do EVPN or even multicast ;-)  so if we're
talking the "smallest viable routing solution", yes, IGP has a higher price
of admission ;-)

--- tony

On Fri, Jan 6, 2017 at 9:37 AM, Juliusz Chroboczek <jch@irif.fr> wrote:

> > You don't give link-state enough kuddos for convergence speed
>
> Ok, here we go.
>
> With a link state algorithm, given enough CPU, convergence speed is almost
> entirely predicated on the speed of LSDB synchronisation.  This gives
> link-state protocols a fundamental advantage as far as convergence speed
> is concerned, at the cost of greater complexity and lesser robustness than
> well-designed distance-vector protocols.
>
> -- Juliusz
>

Re: [babel] babel Digest, Vol 17, Issue 5 Tony Przygienda
Re: [babel] babel Digest, Vol 17, Issue 5 Tony Przygienda
Re: [babel] babel Digest, Vol 17, Issue 5 Juliusz Chroboczek
Re: [babel] babel Digest, Vol 17, Issue 5 Juliusz Chroboczek