Global Security Forum

Security Global Forum

Our mission is to provide clients with an online user community of industry peers and IBM experts, to exchange tips and tricks, best practices, and product knowledge. We hope the information you find here helps you maximize the value of your IBM Security solutions.

 View Only
  • 1.  ISVA 11.0.1 IF1 : after upgrading, refresh tokens are not being invalidated anymore after being used

    Posted yesterday

    Hello,

    We have recently upgraded our UAT ISVA cluster to the latest update (11.0.1 IF1) and thanks to an internal pen-test, we have identified that Refresh tokens are not invalidated anymore after being used.

    I'm able to call the /token endpoint with the same refresh_token as much as I want, with each call generating a new pair of AT/RT/ID_TOKENS.

    -=> Has anyone already noticed this critical change in behavior ? 

    As we are using an OIDC policy where we have not enabled the "Do not rotate refresh token" and "Enable multiple refresh tokens for fault tolerance" this seems more like a critical bug than a desired behavior change.

    I tested the exact same calls on our production environment (still in 11.0.0) and the refresh tokens are correctly discarded after their first use.

    If someone can confirm this, it will be important to advertise that this update (11.0.1 or the IF1) should not go in production for anyone using OIDC/OAuth policies.



    ------------------------------
    André Leruitte
    Security Architect
    POST Luxembourg
    Luxembourg
    ------------------------------


  • 2.  RE: ISVA 11.0.1 IF1 : after upgrading, refresh tokens are not being invalidated anymore after being used

    Posted 13 hours ago

    I saw this come across their alerts back on October 28:

    https://www.ibm.com/mysupport/s/defect/aCIgJ0000006dWXWAY/dt453944?language=en_US&myns=slc&mync=E&cm_sp=slc-_-NULL-_-E

    I suspect it may be related.  I had subscribed to it to monitor for updates, but I don't recall seeing any.  I would definitely report to support to make sure they are aware.

    Your post peaked my interest as I have not tested that firmware version yet, but refresh tokens not expiring would be a big deal for me and our security teams would likely alert on it as well.

    Matt



    ------------------------------
    Matt Jenkins
    ------------------------------



  • 3.  RE: ISVA 11.0.1 IF1 : after upgrading, refresh tokens are not being invalidated anymore after being used

    Posted 13 hours ago

    Hi Matt,

    That known issue is exactly what we have observed. I missed the alert unfortunately so we discovered by chance :( 

    There seems to exist a "test fix" which IBM support will make available as soon as they can.

    Anyway, this is bad news because each time we need to upgrade our ISVA, we do not test 100% of our applications and implementations to make sure something as critical as this has not slipped by. It kind of breaks the trust on the security solution...



    ------------------------------
    André Leruitte
    Security Architect
    POST Luxembourg
    Luxembourg
    ------------------------------



  • 4.  RE: ISVA 11.0.1 IF1 : after upgrading, refresh tokens are not being invalidated anymore after being used

    Posted 13 hours ago

    Glad that was useful.  Over 16+ years using the product, the RTSS (used to be TFIM) is probably one of the largest risks in terms of changes.  I can't imagine how complex that code probably is, especially given they likely rely on a lot of Java libraries.  When you have a lot of customization in the mapping rules, that adds further risk during upgrades.

    At some point, I'd like to develop an automated test harness for all things SAML, OIDC, and OAuth to test the flows and the security around them in my lab before upgrading other environments.  A lot of times we don't think of upgrades, or even our own changes to mapping rules, having such a wide impact.  But when you have countless partners you don't control relying on the infrastructure, it definitely presents a challenge.



    ------------------------------
    Matt Jenkins
    ------------------------------