Thank you for your thoughts. I have done a lot of reading on the topic of CLIENT-CERT/SSL and Tomcat from Google searches as well as the source code of SSLAuthenticator. However I see nothing to indicate how a server would know you from me if I gave you my keystore and keystore password. The server would authenticate you as me. Many tests on my part indicate that is so. Funny that the client does this against the server's cert by default.
I modified a version of SSLAuthenticator to do hostname verification with success, but it seemed to me like something that is either a real omission in Tomcat (perhaps on the basis of adherence to a standards spec.?) or some conceptual misunderstanding on my part. In deference to the thought that I was missing something, I was hoping to hear what I might look at specifically to revert to stock code for SSLAuthenticator. -----Original Message----- From: Bill Barker [mailto:[EMAIL PROTECTED] Sent: Monday, November 08, 2004 12:05 PM To: Tomcat Developers List Subject: Re: org.apache.catalina.authenticator.SSLAutheticator ----- Original Message ----- From: "Ken Sims" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, November 08, 2004 8:25 AM Subject: org.apache.catalina.authenticator.SSLAutheticator >I am curious why SSLAuthenticator does not/cannot compare some >attribute of the client cert with the remote address (requestor)? >Without such a check, it seems to me that certificates are as easily >shared as the credentials used in basic authentication. > There are plenty of tutorials on CLIENT-CERT auth out there. I suggest that you read one :). >Also, why do the realm implementations always return null for >getPrincipal? Couldn't they lookup the user on the users database, >ignoring password, to establish authorized roles? > Usually when you seriously want CLIENT-CERT, you find that you need a custom Realm anyway. There isn't really a one-size-fits-all solution. Where it is implemented in TC (MemoryRealm, and UDBRealm in 5.5.x), it authenticates against the Subject. In a lot of cases, you really only want the CN, or the CN+EMAIL fields. In other cases (e.g. JNDI, JAAS) you might want the CN+entire cert. You can look through BZ for TC 4 to find several examples of proposed implementations. That said, patches are always welcome :). >The combination of these two things seems to me to really limit the >usefulness of client certificate authentication because authentication >provides little guarantee of who the client is, and even if it did, the >client is denied access to any protected resources. > This message is intended only for the use of the person(s) listed above as the intended recipient(s), and may contain information that is PRIVILEGED and CONFIDENTIAL. If you are not an intended recipient, you may not read, copy, or distribute this message or any attachment. If you received this communication in error, please notify us immediately by e-mail and then delete all copies of this message and any attachments. In addition you should be aware that ordinary (unencrypted) e-mail sent through the Internet is not secure. Do not send confidential or sensitive information, such as social security numbers, account numbers, personal identification numbers and passwords, to us via ordinary (unencrypted) e-mail. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
