Re: [TLS] Publication of draft-rhrd-tls-tls13-visibility-00

[Peter Saint-Andre <[e-mail]>]

On 10/22/17 5:26 PM, Steve Fenter wrote:
> I know of a number of large enterprises in verticals including financial, health care, retail, and government, across multiple countries, who are using packet payload inspection within their data centers.  Most of these enterprises are reluctant to step forward in a public forum and reveal their internal network structure and their internal security and monitoring practices. This gives the false impression that out of band decryption of TLS is not a big deal. It is in fact mission critical to a significant number of large enterprises.
> 
> I have been saying to anyone who will listen that the IETF needs a private forum for enterprises, to enable them to come forward and discuss their real requirements. Without this input the IETF is trying to architect and engineer solutions without knowing the complete set of requirements, at least on the enterprise side.  This results in sub-optimal design decisions (from an enterprise perspective), which in this case will break mission critical enterprise monitoring and troubleshooting systems.

The IETF doesn't run private forums behind closed doors. You'd need to
do that kind of work elsewhere (these large enterprises could, of
course, start their own industry forum, where they could work in ways
that the IETF doesn't).

> We've already experienced what a rollout of TLS 1.3 will be like, at more than one enterprise, when certain vendors decided to move Diffie Hellman ciphers to the top of their priority list on a code upgrade. This caused severity one outages of critical monitoring systems. 

It sounds as if different internal teams might not have been
communicating well about the rollout of those new cipher suites.
Operational issues in large enterprises are not a problem that requires
protocol work.

>  This means that critical applications depend on these monitoring systems, and if the monitoring system is down the application is completely down. This is not the outcome we want when TLS 1.3 is rolled out, but it is what we are headed for. Enterprise monitoring should be tested as part of the operational TLS 1.3 testing before TLS 1.3 is approved as a standard, and TLS 1.3 should not be approved if enterprise monitoring breaks.

Operational testing is always good, but very strong arguments need to be
made for the latter claim. Among other things, you're adding a new
requirement to the Internet Standards Process, which would necessitate
IETF consensus on changes to RFC 2026!

> The only other option being presented to enterprises is that we continue to run on a TLS spec that is nine years old, and then continue running it until it is 14 to 19 years old. It makes no sense to me to put out a TLS 1.3 standard, but say that enterprises cannot upgrade to it.

There are many options, some of which Kathleen outlined in her blog
post. It's not helpful to say there is just one option when we haven't
fully explored either the problem space or the solution space. And by
"we" I mean especially those who are claiming the need for TLS visibility.

Peter