[tz] [Tzdist-bis] [calsify] tzdist and IANA -- estimating the operational parameters

[Eliot Lear]

Hi Steve,


> On 18 Jul 2019, at 15:50, Steve Crocker <steve at shinkuro.com> wrote:
> 
> Early in this thread it was mentioned the the time zone database should be served in a fashion similar to DNS.  My first thought was the numbers are wildly different.  I jotted down a first cut at identifying the relevant operational parameters.  Perhaps the people proposing this service can flesh out the quantitative aspects.


My issue isn’t so much the size of entries.  Many entries can fit in a DNS packet *without* EDNS0.  It wouldn’t be wildly crazy to to use 2k as a good size.  The problem is 2k*O(10^9) (and soon O(10^10)) devices X an unknown frequency, because it is in part based on political considerations of the various jurisdictions.  The chaos coming to Europe soon will demonstrate this.  Also, I don’t think we should swear by the code in devices these days in terms of how they back off or what epsilons they use in their timers.

And then there are all the security aspects of that.  Shoving this stuff into the DNS would pretty much require DOH everywhere.

Eliot

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: Message signed with OpenPGP
URL: <http://mm.icann.org/pipermail/tz/attachments/20190718/9cba5320/signature.asc>