For more information regarding the security incident at F5, the actions we are taking to address it, and our ongoing efforts to protect our customers, click here.

Forum Discussion

newf5learner_13's avatar
newf5learner_13
Icon for Nimbostratus rankNimbostratus
Feb 22, 2017

F5 ASM deployment for production traffic in transparent mode.

Hi Experts,

 

I'm new to the F5 ASM deployment, so posting here with my concerns.

 

Background information: I have enabled ASM for the VIPs which are already in production. I don't have any trusted testers, so I don't have chance to get some Trusted IP addresses that can be added. As its a production deployment, I chose the Transparent Mode with 7 days enforcement period with Automatic policy builder feature.

 

Application information: Tomcat based apps with SQL DB and with SSL enabled.

 

I'm seeing manual traffic learning stats as shown below..

 

 

Also, I could see the entries for under Enforcement readiness, there are too many signature suggestions.

 

 

Could someone suggest me what is the next I should do? I'm seeing many other parameter violations and URL length, parameter input violations etc.

 

I'm trying make this policy as the bench mark for any other policies I'm going to enable. As this is my first policy, I would like to be careful with the an changes to it as I want to export this policy for Blocking mode enforcement for other production VIPs.

 

Please help me with your suggestions.

 

thanks.

 

7 Replies

  • Hi, First of all, do you think 7 day enforcement is enough for traffic learning? We usually go for at least 14 days to a month. You need sufficient amount of traffic to get ASM learn the traffic behavior. Once we have enough traffic and violation events perform the violation analysis one-by-one. You need to work with Apps team and challenge them about these violations, many cases you need to remediate the false positives and some other cases you need to challenge them. I think you also should decide on positive and negative security models you planning to go ahead. We usually go this flow of violations

     

    1) RFC compliance 2) Length Limits 3) Valid File Types 4) Valid URLs 5) Valid Parameters 6) Parameter Compliance 7) Valid meta characters 8) HTTP Method / Headers 9 Attack Signatures / Virus

     

    hope this helps

     

    • newf5learner_13's avatar
      newf5learner_13
      Icon for Nimbostratus rankNimbostratus

      thanks for the reply.

       

      What are the negative security violations in the traffic learning?

       

      Can I accept the suggestions the policy builder listed above? Can I consider the suggestions are the valid ones to be accepted in the policy and so that they will not be triggered as an event further?

       

    • Amine_Kadimi's avatar
      Amine_Kadimi
      Icon for MVP rankMVP

      I'm not sure exporting an application security policy and reusing it for other VIPs will bring any advantage, as most applications don't share the same components and don't work similarly, each specific application will have its own urls, parameters, cookies, and file types that the ASM will have to know, this is positive approache. In the negative approach, each application will need each own sets of attack signatures targeting the specific components of the applications (OS, database, php, asp...).

       

    • newf5learner_13's avatar
      newf5learner_13
      Icon for Nimbostratus rankNimbostratus

      Thanks Amine and Vijith

       

      I read some article where it was stated that F5 recommends creating a baseline policy for the environment which includes basic security requirements that are embedded into the policy components. And such policy can be considered as template which can be re-used. So, was thinking if I can make a standard policy and try replicating it in transparent mode with enforcement period of 7 days for all the other VIPs.

       

      could you please let me know what are the negative security violations in the traffic learning? Should we allow them or enforce them?

       

  • To add what Amine already commented, You should always consider a balanced approach for positive and negative models. Just allowing signature (negative model) will not help you to shut out all the malicious traffic, so you should also think on the positive model (ASM learns many stuffs) for zero day and all.

     

    Regarding baseline, yes we can have a baseline for all types of application and start from modify the parameters for specific applications. this what exactly we do especially you are dealing with many different types of applications. cheers