On Fri, Sep 25, 2015 at 07:40:05PM +0100, Jeremy Harris wrote:
> Why is it not possible for TLS1.3 to provide that same service
> combination, but implemented by design in a layered fashion?
TLS is correctly agnostic of semantic boundaries, in application
data. For this to work, applications would need to be able to ask
TLS to enable and disable compression at any time after the handshake,
once some uncompressed or compressed data has gone by.
This requires new application protocol verbs "STARTCOMPRESSION",
"STOPCOMPRESSION", and underlying support in the TLS layer.
A suitable application I/O library that supports pushing and popping
protocol "modules" onto a raw I/O stream, would be a better vehicle
for this than forcing dynamic compression support into TLS.
--
Viktor.
_______________________________________________
TLS mailing list
TLS@ietf.org
https://www.ietf.org/mailman/listinfo/tls
