Re: [Idr] draft-snijders-idr-rfc8203bis

[Jared Mauch <[e-mail]>]

Job,


> On Dec 8, 2017, at 10:47 AM, Job Snijders <[e-mail]> wrote:
> 
> Dear IDR,
> 
> So, it appears I was wrong to push for the arbitrary maximum length of
> 128 octets when we worked on the BGP Administrative Shutdown
> Communication (RFC 8203). At the time, multiple people pointed out that
> it might be better to let the the thing be a little bit longer (255),
> and in retrospect that indeed would've been better. I'm sorry :(
> 
> At this month's peering forum meeting of the Moscow Internet Exchange,
> Misha Grishin (MSK-IX) reported on experiences with the shutdown
> communication, and indicated that the concept was considered useful, but
> that 128 octets is too small in context of multibyte character sets such
> as Cyrillic script and they couldn't easily fit the short messages they
> wanted to send in the communication. The GoBGP authors also questioned
> the 128 limit during the implementation phase, I now see why.
> Operational feedback like this is invaluable and worth our
> consideration.

I support fixing this.

Is 255 the right number?  If all the chars are 4 byte that supports 63 chars.

I’m by no means a unicode expert, but if we fix it, we should ensure the right
number is used.

- Jared

ps - wanting to avoid future 💩