[Tzdist] Security Considerations: Is TLS security sufficient

Daniel Migault <mglt.ietf@gmail.com> Thu, 21 August 2014 15:53 UTC

MIME-Version: 1.0
Date: Thu, 21 Aug 2014 17:52:54 +0200
Message-ID: <CADZyTkkACqAC9f4eTcUPhAqGLgLOhB+CegLDdutcNExq0Bfe=w@mail.gmail.com>
From: Daniel Migault <mglt.ietf@gmail.com>
To: tzdist@ietf.org
Content-Type: multipart/alternative; boundary="bcaec53d57afb144e5050125b6ea"
Archived-At: http://mailarchive.ietf.org/arch/msg/tzdist/R9q9cPY1MypkLU6z-ZMU8MmU2DY
Subject: [Tzdist] Security Considerations: Is TLS security sufficient
Precedence: list

Hi,

The draft mentions that security is provided by TLS. This model is fine as
long as I trust the server I am connected to. However, if we consider the
root provider as legitimate (IANA for example), a TLS communication would
provide integrity check. On the other hand, we have no mean to check that
the alternate providers have not alterate/modified the time zone data.

Wouldn't it be more appropriated to have integrity check mechanisms within
the data.

The charter mentions

  - The time zone data will be based on the Time Zone Database
  (http://www.iana.org/time-zones) but must be able to include any source
  of time zone data.

So working on the data format seems out of scope. However if a secure data
format may be designed in the future. TLS will not be needed for integrity
protection. but only for confidentiality or privacy issues. Similarly to
DNS, it would make sense to have both secure or unsecured transport layer
depending on your goals.

-- 
Daniel Migault
Orange Labs -- Security
+33 6 70 72 69 58

Re: [Tzdist] Security Considerations: Is TLS secu… Cyrus Daboo
[Tzdist] Security Considerations: Is TLS security… Daniel Migault