eap Mailing List: Re: draft-urien-eap-smartcard-06.txt

[Pascal Urien (Pascal.Urien]

Hi Thomas

At 18:00 20/08/2004 +0200, Thomas Otto wrote:
> Hi Pascal,
>
> I just had a look at your website ...
> Under the headline "the first EAP-TLS smartcard is operational
> in the ENST Wi-Fi network" you provide an ethereal packet dump of
> a full EAP-TLS authentication.
> I measured the total time of this conversation. The time difference between
> EAP-Request/Identity at 17:46:21 and EAP-Success at 17:46:58, are
> 37 seconds.

That's correct.


> More precisely, the client needs almost 30 seconds to send
> ClientKeyExchange, CertificateVerify, and ChangeCipherSpec
> (in frame 52), where I remember ClientKeyExchange requires
> the 48 byte premaster secret. (which requires expensive
> computations ..)

This step cost around 1x RC4, 3x  RSA-1024 and about
 1000 x MD5&SHA1 512bits- blocks computation.

Most of the time is spent in digests functions


> So, is there hope for a speedup or is EAP-TLS not suitable for the
> EAP smartcard? How much Java influences this duration?

Usually digests functions are not supported by crypto processors
and are directly computed by 8 bit micro-controllers, that why they
are slow.

Some crypto algorithms like RC4, HMAC-MD5, HMAC-SHA1, TLS-PRF
are written in java card language. The solutions in order to dramatically
decrease the computing time are the following

- using smart cards with crypto processors supporting digests functions
- defining java card API for HMAC-MD5, HMAC-SHA1 and TLS-PRF
  (in progress ...)

> In contrast to this, as answer to "Are smartcards performances
> sufficient ?" we get the information, "Usually smart
> cards include crypto-processors that compute the RSA 2048 bits
> algorithm in less than 0,5s."

All crypto processors supports RSA. RSA is not an issue for
smart cards.

> This is great, so I suppose the card above lacks of such a
> crypto-processor ?

As i mentioned it above, RSA is not an issue. Digests functions
are the main bottleneck performances


> May I compare this to the running time of my implementation of
> EAP-PSK, which took for a full authentication only 0.75 seconds.
> There is a strange 0.25 second delay between Identity Request and
> Response, so essentially we have 0,5 seconds for the protocol.
> By the way, the notebook used for this test was nothing special, just a
> 4year old 400 MHz machine ;-)


Smartcard is usually a 10 MHz 8bit micro-controller including a 
crypto-processor..

AES is supported by javacard 2;2. I have not yet performances figures for
EAP-PSK, it seems compatible with existing components.


> Thomas


Pascal


> References
> [1] http://www.infres.enst.fr/~urien/security/eap-tls-trace.pdf
> [2] http://t13.mine.nu/EAP-PSK/020604-eappsk.pcap
> _______________________________________________
> eap mailing list
> eap [at] frascone.com
> http://mail.frascone.com/mailman/listinfo/eap