Thanks for the link to the performance paper. This paper has some
wrong statements worth being noted:

Chapter 5.3, item 3 states:
"Encrypt the whole shebang (UsernameToken + timestamp + nonce) with the
RSA
private key of the client."

To be precise here: WSS uses a symmetrical cipher to encrypt the 
"whole shebang", WSS generates a random key for this symmetrical cipher
and encrypts only this random key with RSA


Chapter 6, Figure 1:
The description of figure 1 is not correct with respect to "mod 2". The
document states:
"Mod 2: UsernameToken + timestamp + nonce together encrypted with
the RSA client's private key;"

This is wrong. The client uses the server's public key to encrypt the
random key used to encrypt the "whole shebang" :-) . This has several
impacts
to the shown figures and explains the big difference between mod2 and
mod3:
- encrypting with a public key is much faster then encrypting with a
  private key (this is an attribute of most public key algos)
- decryption in RSA even with the private key is faster than encryption

After all, mod2 is "moderate" in terms of encryption/decryption
overhead.
Mod 3 uses Signature which implies RSA encryption of the signature hash
with the client's private key - which is time consuming. This explains
the
big difference between the two modes


Chapter 7, point 1 states:
"One should completely free the security package from the subordinate
Web service
engine. For example wss4j works only with the Axis Web service engine.
This is
understandable from a developer's point of view but a poor design
choice"

This is completely wrong - the WSS4J security library is independent of
Axis (at that point in time Axis 1). Only the handlers are dependent of
the
unerlyin SOAP engine - WSS4J included a handler for Axis-1 and a handler
according the Java JCP standards, independent of Axis-1. Also Axis-2
uses WSS4J - and Axis-2 has a complete different implemetation how to
integrate "handlers" (interceptors) and how to deploy.

Regards,
Werner


> -----Original Message-----
> From: ext Ostermueller, Erik [mailto:[EMAIL PROTECTED] 
> Sent: Monday, October 13, 2008 8:23 PM
> To: jimmy Zhang; Werner Dittmann
> Cc: [EMAIL PROTECTED]; [email protected]
> Subject: RE: WSS4J 1.5.4 Encryption Performance Question
> 
> This also might be helpful:
> http://lis.fh-aargau.ch/nicola/pub/WSS4Jperformance.pdf 
> 
> -----Original Message-----
> From: jimmy Zhang [mailto:[EMAIL PROTECTED] 
> Sent: Monday, October 13, 2008 12:53 PM
> To: Werner Dittmann
> Cc: [EMAIL PROTECTED]; [email protected]
> Subject: Re: WSS4J 1.5.4 Encryption Performance Question
> 
> http://www.w3.org/2007/xmlsec/ws/papers/06-zhang-ximpleware/
> ----- Original Message -----
> From: "Werner Dittmann" <[EMAIL PROTECTED]>
> To: "jimmy Zhang" <[EMAIL PROTECTED]>
> Cc: <[EMAIL PROTECTED]>; <[email protected]>
> Sent: Saturday, October 11, 2008 1:19 AM
> Subject: Re: WSS4J 1.5.4 Encryption Performance Question
> 
> 
<SNIP ----- SNAP>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to