Application Whitelist Example

Keep in mind that you do not need to capture every application that might be in use on your network in your initial inventory. Instead you should focus here on the applications (and general types of applications) that you want to allow. Temporary rules in the best practice rulebase will catch any additional applications that may be in use on your network so that you are not inundated with complaints of broken applications during your transition to application-based policy. The following is an example application whitelist for an enterprise gateway deployment.
Application Type
Best Practice for Securing
Sanctioned Applications
These are the applications that your IT department administers specifically for business use within your organization or to provide infrastructure for your network and applications. For example, in an internet gateway deployment these applications fall into the following categories:
  • Infrastructure Applications—These are the applications that you must allow to enable networking and security, such as ping, NTP, SMTP, and DNS.
  • IT Sanctioned Applications—These are the applications that you provision and administer for your users. These fall into two categories:
    • IT Sanctioned On-Premise Applications—These are the applications you install and host in your data center for business use. With IT sanctioned on-premise applications, the application infrastructure and the data reside on enterprise-owned equipment. Examples include Microsoft Exchange and active sync, as well as authentication tools such as Kerberos and LDAP.
    • IT Sanctioned SaaS Applications—SaaS applications are those where the software and infrastructure are owned and managed by the application service provider, but where you retain full control of the data, including who can create, access, share, and transfer it (for example, Salesforce, Box, and GitHub).
  • Administrative Applications—These are applications that only a specific group of administrative users should have access to in order to administer applications and support users (for example, remote desktop applications).
General Types of Applications
Besides the applications you officially sanction and deploy, you will also want to allow your users to safely use other types of applications:
  • General Business Applications—For example, allow access to software updates, and web services, such as WebEx, Adobe online services, and Evernote.
  • Personal Applications—For example, you may want to allow your users to browse the web or safely use web-based mail, instant messaging, or social networking applications.
The recommended approach here is to begin with wide application filters so you can gain an understanding of what applications are in use on your network. You can then decide how much risk you are willing to assume and begin to pare down the application whitelist. For example, suppose you find that Box, Dropbox, and Office 365 file-sharing applications are all on use on your network. Each of these applications has an inherent risk associated with it, from data leakage to risks associated with transfer of malware-infected files. The best approach would be to officially sanction a single file-sharing application and then begin to phase out the others by slowly transitioning from an allow policy to an alert policy, and finally, after giving users ample warning, a block policy for all file sharing applications except the one you choose to sanction. In this case, you might also choose to enable a small group of users to continue using an additional file-sharing application as needed to perform job functions with partners.
Custom Applications Specific to Your Environment
If you have proprietary applications on your network or applications that you run on non-standard ports, it is a best practice to create custom applications for each of them. This way you can allow the application as a sanctioned application and lock it down to its default port. Otherwise you would either have to open up additional ports (for applications running on non-standard ports), or allow unknown traffic (for proprietary applications), neither of which are recommended in a best practice Security policy.

Related Documentation