Re: [tcpinc] Benoit Claise's No Objection on draft-ietf-tcpinc-tcpcrypt-09: (with COMMENT)

[Daniel B Giffin <[e-mail]>]

Benoit Claise wrote:
> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
> 
> Nothing against the publication of this document but ... as for any
> experimental RFCs, we must describe the criteria for a successful experiment
> (evaluation)?

Oh right!

I've added this section:

	9.  Experiments

   Some experience will be required to determine whether the tcpcrypt
   protocol can be deployed safely and successfully across the diverse
   environments of the global internet.

   Safety means that TCP implementations that support tcpcrypt are able
   to communicate reliably in all the same settings as they would
   without tcpcrypt.  As described in [I-D.ietf-tcpinc-tcpeno]
   Section 9, this property can be subverted if middleboxes strip ENO
   options from non-SYN segments after allowing them in SYN segments; or
   if the particular communication patterns of tcpcrypt offend the
   policies of middleboxes doing deep-packet-inspection.

   Success, in addition to safety, means that hosts which implement
   tcpcrypt actually enable encryption when they connect to each other.
   This property depends on the network's treatment of the TCP-ENO
   handshake, and can be subverted if middleboxes merely strip unknown
   TCP options or if they terminate TCP connections and relay data back
   and forth unencrypted.

   Ease of implementation will be a further challenge to deployment.
   Because tcpcrypt requires encryption operations on frames that may
   span TCP segments, kernel implementations are forced to buffer
   segments in different ways than are necessary for plain TCP.  More
   implementation experience will show how much additional code
   complexity is required in various operating systems, and what kind of
   performance effects can be expected.