Use a web application firewall (WAF)

A web application firewall (WAF) is an application layer protection for bi-directional web-based traffic. With a WAF, you can track web traffic and use specific tools to configure access control for your web content. Doing this improves your service’s security monitoring and security position.

Why you should use a WAF

Your continuous integration (CI) and continuous deployment (CD) pipelines should include security tests in their workflows to identify any common vulnerabilities in your code. Some common vulnerabilities like Cross-site Scripting (XSS) and XML command injection attacks are still possible in your production environments due to human error.

Combining a WAF with CI and CD tools reduces the risk from those tools, and provide enhanced layered security coverage for your service.

You may also need to use a WAF because of:

GDS or departmental policies or standards
Information Assurance (IA) requirements to minimise risk
Payment Card Industry Data Security Standard (PCI DSS) compliance

When and how to use a WAF

Set up a baseline of tests in your project’s alpha phase to identify any security vulnerabilities. As your service’s features grow, extend your tests to cover new vulnerabilities you identify. For example, through exercises like application threat modelling

Good development practices should detect and fix common vulnerabilities before they reach production environments. Use your WAF to track digital services vulnerabilities an attacker could exploit.

You should:

have an independent security audit in place
use established logging techniques
encrypt data at rest as well as in transit
subscribe to and apply security patches
use query variables instead of plain text (stored procedure) to prevent SQL injections

You should monitor the Open Web Application Security Project (OWASP) top 10 most critical web application vulnerabilities to keep up to date with the latest threats.

Using a WAF should align with your other security monitoring features, like Security Information Event Management (SIEM). When developing use cases you should also factor in the extra time and resources needed to configure WAF rules.

Managing your WAF

Identify any areas in your app not covered by your WAF and define measures to protect them, such as using:

alerts - as part of your incident management strategy
threat modelling - to assess potential weaknesses in your environment
reviews - to manage your security controls

Alerts

It’s not always possible to block a detected attack because some services need to process transactions for any user of that service. In these situations you should raise events as alerts.

When WAF alerts are raised, make sure you already have an incident policy in place, including:

who should manage an alert
what their responsibilities are
how to investigate an alert
how to tune out false positives

Reviews

Review your WAF after each application change against the risks in the OWASP top 10 category rules.

This should be similar to how you use an IT Health Check (ITHC) to test and confirm the effectiveness of security controls in your environment.

Case study GOV.UK PaaS

A GOV.UK PaaS tenant uses a pattern with Amazon Web Services (AWS) WAF before forwarding traffic to their apps with enabled shield advance for extra protection.

For more information read the proposed architecture for implementing a DDoS-resistant Website using AWS Services.

Case study GOV.UK Pay

GOV.UK Pay uses a NAXSI WAF (pay-nginx-proxy) for its Nginx, forked from the Home Office.

GOV.UK Pay operates under the governance of PCI compliance and DSS point 6.6 which states the need for web application scanning.

Contact Cyber Security

Contact the security architects in the GDS Cyber Security team by email at cyber.security@digital.cabinet-office.gov.uk or use the #cyber-security-help Slack channel for help and advice.