Re: [TLS] draft-mavrogiannopoulos-new-tls-padding-00

[Nikos Mavrogiannopoulos <[e-mail]>]

On Fri, 2013-11-29 at 22:45 +0100, Martin Rex wrote:

> > http://tools.ietf.org/html/draft-mavrogiannopoulos-new-tls-padding-00
>  
> If you leave the fixing (and changing) of the Cipher PDUs alone,
> you could alternatively pursue a "random padding" by simply defining
> an (anti-)compression algorithm to that effect, with whatever semantics you
> desire.  This might also fit smoother with implementations.
> 
> The hard limit that your fancy (anti-)compression algorithm will
> have to observe from the TLS specs,
>   e.g. http://tools.ietf.org/html/rfc5246#section-6.2.2
> 
> is that the output of your (anti-)compression must not exceed 2^14+1024
> bytes.
> The "may not increase the content length by more than 1024 bytes" rule
> is more of a soft rule that can be adjusted by the (anti-)compression
> algorithm spec as long as the "the output must not exceed 2^14+1024"
> hard rule is obeyed.

That's a pretty clever observation, and would be of quite use if the new
padding mechanism is not used. Note however, that one would want to keep
a uniform size in records and padding should be within the 2^14 limit
(or the max data limit); otherwise it may be of little use in hiding the
length.

regards,
Nikos