AWS Security Blog

How to build a Security Guardians program to distribute security ownership

Welcome to the second post in our series on Security Guardians, a mechanism to distribute security ownership at Amazon Web Services (AWS) that trains, develops, and empowers builder teams to make security decisions about the software that they create. In the previous post, you learned the importance of building a culture of security ownership to scale security within your organization, and how AWS achieves this using the Security Guardians program. Since then, many customers have asked how they can build their own, similar program.

In this post, you will learn the steps to build your own Security Guardians program for your organization, including how to:

  • Set the vision, mission, and goals of your program
  • Identify developer teams that can pilot your new program
  • Define the expected behaviors for those teams
  • Develop training and create opportunities for career development to keep your teams engaged in the program

The guidance in this post is based on what we learned at AWS. Because every organization is different, the final version of the program you build is likely to look different from the one at AWS. Your program needs to reflect the current state of your organization’s culture of security and be designed to cultivate the security-related behaviors that are most important to your organization.

Security Guardians program mechanism

As discussed in the previous post, mechanisms form a key part of our business at AWS. Figure 1 demonstrates how a mechanism is a complete process, or virtuous cycle, that reinforces and improves itself as it operates. It takes controllable inputs and transforms them into ongoing outputs to address a recurring business challenge. In this case, the business challenge AWS faced was that security findings were being identified late in the development lifecycle, making it more expensive—in terms of time, money and effort—to remediate them. This led to bottlenecks in our security review process. The culture of security at AWS, specifically our culture of ownership, provides support to solve this challenge, but we needed the Security Guardians mechanism to actually do it.

Figure 1: AWS mechanism cycle

Figure 1: AWS mechanism cycle

With most mechanisms, driving adoption is difficult, especially when the mechanism requires human participation to succeed. This is also true in the case of Security Guardians, and you can use our experience to help you avoid some of the challenges and growing pains of driving adoption.

Getting everyone aligned

“If I had an hour to solve a problem and my life depended on the solution, I would spend the first 55 minutes determining the proper question to ask, for once I know the proper question, I could solve the problem in less than five minutes.” – Albert Einstein

Getting alignment for the need to distribute security expertise starts with deeply understanding what problems need to be addressed. For example:

  • Is product delivery velocity being negatively impacted by delays in the security review process?
  • What business goal or metric are these delays negatively impacting?
  • Where in the security review process are those delays occurring?
  • What factors are contributing to those delays?
  • Is it a lack of time, people, or skills?

Thoroughly understanding the specific problems and their root causes, as identified by answering those questions, allows you to evaluate whether distributing security ownership is the appropriate solution. This in turn makes it easier to gain alignment and buy-in across the organization for the chosen approach.

A component of a culture of security

Building a strong culture of security requires support from executive leadership to set the direction for the rest of the organization. Executive support makes it easier for product leaders to secure the resources and finances needed for a Security Guardians program to be successful. To align with your organization’s leaders, you can reflect on the goals of your leaders and how the Security Guardians program can be built to meet those goals.

For example, if your business goal is to ship products 25 percent faster, understand how a particular resourcing effort from Security Guardians is going to help your organization meet that goal. AWS benefited from the program with a 26.9 percent reduction in the time to review a new service or feature when a Security Guardian was involved.

Our experience is that it’s challenging to establish a Security Guardians program without executive support. If you’re struggling to identify a business leader to sponsor the program and provide insight on the business problem, your AWS account team—including your account manager or solutions architect—can help. If you’re a business leader or executive reading this post, consider becoming that sponsor yourself.

One step at a time

A step-by-step approach to implementing the Security Guardians program helps overcome organizational challenges and avoid common pitfalls that could lead to failure. These steps, shown in Figure 2, are:

  1. Set the vision
  2. Choose innovators
  3. Define behaviors
  4. Maintain interest
  5. Measure success

These steps support the activities that make a mechanism successful: adoption, inspection, and tools.

Figure 2: Steps for implementing a Security Guardians program

Figure 2: Steps for implementing a Security Guardians program

Set the vision

Now that you’ve identified the business problem or business goal, set the vision for the Security Guardians program by working backwards from this problem or goal to define the purpose of your program. For example, the vision of the AWS Security Guardians is “To nourish security ownership that consistently delights our customers with security-by-design throughout the development lifecycle.”

Craft an ambitious vision for your Security Guardians program. Think beyond easy wins and focus on bold, forward-thinking security outcomes for your organization. Make sure that each element of your vision aligns with a business problem or goal. The following table is an example of how the vision of the program is aligned with business goals:

Business goals Security outcome Long-term goals
Develop products faster and more efficiently. To improve developer agility while reducing security risk. Increase the number of threat models performed by Security Guardians (instead of by application security engineers). Over time, this goal could change to “increase the quality of threat models.”.

Decrease the average monthly security issue rate.

Train three new Security Guardians each quarter.

Reduce long-term security spend. To identify and mitigate security risk as early as possible.
Increase customer trust. To exceed customer security expectations by raising the security bar.

The next step is to define a clear mission that is supported with measurable goals. The mission and goals must be achievable and help to move the needle towards the long-term vision.

The final part is to name your program. We chose Security Guardians, like Marvel’s Guardians of the Galaxy. We’ve also heard customers using Security Champions, Security Advocates, Security Innovators, and Security Drivers. Have fun with it and make sure the name resonates with as many participants as possible.

After you’ve defined the vision, future state, mission, measurable goals, and name of the program, review them with your security and business leaders. It’s beneficial to include your innovators or Security Guardians who will be early adopters of the program in this review. In the next section, you’ll learn how to identify these innovators.

Choosing innovators

Just as you develop for and iterate with early adopters of the products you’re building, you should identify individuals and teams who will pilot the program with you. Before the AWS Security Guardians program, our application security engineering teams built relationships with product teams through security reviews.

This meant that they already knew which individuals within those product teams had an interest in security. This is where AWS started, but the success of your program isn’t dependent on whether you already know who these individuals are. Development teams will self-identify and nominate Security Guardians from their own teams. Figure 4 shows examples to help you get started understanding which development teams will be good early adopters for your program.

Figure 3: Example product teams for early program adopters

Figure 3: Example product teams for early program adopters

The examples in Figure 3 include:

Candidate A: Quick wins team

Early adopters typically share key traits, including existing security measures and a designated security role or team members with security expertise. Essentially, they already prioritize security at the team level.

Candidate B: High impact team

This is the team most impacted by the disparity between product development teams and security teams; the agility and time-related benefits of the Security Guardians program will be the highest for this team. For example, this team might be facing long delays in launching products because of the current security review process at your organization.

Candidate C: High risk team

This team owns a product that has a high security risk because of the nature of the product. This team will benefit the most from additional security scrutiny and from raising the security bar at your organization. For example, this team might be building a product that’s considered a critical asset, hosts sensitive data, or performs critical processes.

After you’ve identified one or more teams that could be good early adopters of the program, you need to identify at least one individual from each team to serve as the Security Guardian. Keep the vision and goals of your program in mind when selecting your Security Guardian. Your early Security Guardians should have at least the following characteristics:

  • Ability to exercise well-informed and decisive judgement
  • Maintain and showcase their knowledge
  • Not afraid to have their work be independently validated
  • Advocate for their security needs in internal discussions
  • Hold a high security bar
  • Thoughtful and assertive to make customer security a top priority on their team

In terms of time commitment, our experience is that each Security Guardian spends an average of 3.5 hours each month on activities such as answering general security questions, identifying security stories needed for sprints, diving deep into security related tasks and supporting security related tasks. Each application security review takes approximately 4 hours of effort.

The first post of this series contains even more details on the characteristics that make a good Security Guardian.

Defining behaviors

It’s important to set expectations on what behaviors you want Security Guardians, developers, and security teams to exhibit within the context of the program. These behaviors typically relate directly to the goals of the program. For example, if one of the goals is to increase the number of threat models created, then create threat modeling will be one of the defined behaviors. The behaviors need to be measurable with some flexibility for change as you improve the program.

At AWS, our Security Guardians have access to a runbook that lists the activities each Guardian should take when engaged as part of a review. With each of these activities understood, the program team will then make sure appropriate training is provided so that the Security Guardians are able to complete each of the activities. For example, AWS Security Guardians are asked to help develop threat models. To support this, the program team has developed and released training material to teach Security Guardians how to create a threat model.

With the defined behaviors, understand how the Security Guardian and product development team will engage with the security team. Although we’re clearly defining behaviors, the behaviors aren’t typically done in a silo for the successful launch of a secure product. At AWS, the Security Guardians and product developers engage with the security teams in key partnership areas. If you’re unsure of where to start in defining the behaviors of your program, Figure 4 shows an example of how teams interact at AWS, beginning with the creation of an initial threat model and going through review, remediation, and testing. Consider creating your own version of the model to help define the behaviors and key partnership areas at your organization.

Figure 4: Example behaviors and partnership areas at AWS

Figure 4: Example behaviors and partnership areas at AWS

In the example of a threat model review, the Guardian and the central security team will jointly create and review the threat model. Specific activity examples include reviewing threats that have no documented mitigations and discussing additional threats that haven’t yet been considered.

As part of encouraging a culture of ownership, AWS recommends allowing Security Guardians to influence the role within a set of boundaries. An example of this is allowing the Security Guardians to be a part of recurring reviews of the program growth metrics, actively collecting their feedback, and encouraging them to host their own training sessions. Active Security Guardians are key to the success of the program and allowing them to influence the program will give them a sense of ownership and inclusion.

Maintaining interest

It’s important to not lose sight that a program like the AWS Security Guardians program is supported by volunteers. Most of your Security Guardians will be product developers who already have a full-time job developing products for your organization. The time and effort to find and onboard new Security Guardians will have a low return on investment if they stop engaging because the program owners didn’t keep them engaged. Keeping Security Guardians is just as important as finding them.

At AWS, we invest time to understand how to build trust with Security Guardians and provide value by working backwards from their wants and needs. Some Security Guardians joined the program to learn new skills and for career growth opportunities. AWS built training programs that were designed for Security Guardians and provide metrics that are used to document their impact to their managers and leaders.

AWS Security Guardians constantly tell us that they value recognition of their contributions by leadership. We work to build mechanisms to continuously surface the great work of our Security Guardians. We also recognize the contributions Security Guardians make through awards, gifts, and other incentives. For example, each quarter, the AWS Security Guardians team sends out a newsletter to senior leaders of the organization. This communication identifies the Guardians within their organization and highlights their contributions, including the number and impact of reviews they’ve completed.

Another way that AWS recognizes the contributions of our Security Guardians is through the Guardians Belt Program. The Guardians Belt Program is designed to recognize Security Guardians for their contributions and support them as they work to advance their security skills and expand their scope of impact. Security Guardians earn Black, Green, Yellow, and White belts with each belt corresponding to significant accomplishments that require consistent commitment to raising the security bar.

To make sure that Security Guardians value the program, your organization should provide and actively facilitate benefits. The benefits must be accessible without requiring additional time or effort from the Security Guardians, promoting immediate and direct gains. Consider the following examples of benefits to maintain Security Guardian interest and support:

  • Specialized training: Workshops, game days, challenges and contests.
  • Impact opportunities: Ability to impact multiple products by working with other teams in the organization, ability to help define patterns, best practices, and automation for the program.
  • Community: Collaborate, connect, share and learn from experts and individuals with similar interests.
  • Ownership opportunities: Ability to accelerate certain steps in the process.
  • Leadership opportunities: Active involvement in recurring program or business reviews.

The best ways to maintain interest are determined by the culture of your organization. What does your organization value the most, and how will the program provide that to your Security Guardians? Sometimes, the best way to answer these questions is to ask your early or potential Security Guardians.

Measuring success

The final step of building a successful Security Guardians program is to measure program success. Measuring success is equivalent to the inspection step from Figure 1. This verifies that your desired outcomes are being achieved and provides a jumping off point for iteration. Measuring success also gives you the opportunity to audit the output or results of the Security Guardians program and perform corrections and improvements.

Earlier in this post, we covered identifying the business problem and creating the vision and measurable goals for your Security Guardians program. Example metrics include:

  • Average time to release features
  • Average number of security issues per team
  • Average time spent by Security Guardians and builders doing security work
  • Percentage of Security Guardians who have taken required and non-required training

Measuring success includes steps to collect feedback and tune the program over time, shown in Figure 5.

Figure 5: Feedback and tuning steps for Security Guardians program.

Figure 5: Feedback and tuning steps for Security Guardians program.

The cycle to gather feedback and tune the program includes:

  1. Report on metrics
  2. Communicate wins
  3. Measure outcome and cycle time
  4. Identify trends
  5. Review goals

Gathering feedback from Security Guardians is as important as providing feedback to them. One of the ways AWS collects feedback from Security Guardians is through an annual survey that collects feedback on their experiences of program and tooling. To help both builders and Security Guardians improve over time, our security review tooling captures feedback from security engineers on the inputs from Security Guardians. Combined, the data gathered through these surveys helps our security ownership mechanism reinforce and improve itself over time.

Figure 6 summarizes the steps that you can take to develop your program.

Figure 6: Security Guardians program steps

Figure 6: Security Guardians program steps

The broad steps to develop a program include:

  • Set the vision: Set your vision for the program and metrics for success. Get sponsorship from leadership. Choose a name for your program.
  • Choose innovators: Identify innovators who have a passion for security and foster a community with continuous knowledge sharing.
  • Define behaviors: Redefine your RACI (responsible, accountable, consulted, informed) and be clear on expectations from your security advocates.
  • Maintain interest: Provide clear training and learning paths and opportunities for career advancement.
  • Measure success: Gather feedback and measure the program’s effectiveness.

Conclusion

This post and the previous post covered numerous concepts, considerations, and ideas, including:

  • The initial intention of the Security Guardians program is to focus on training developers in product teams. This improves early security-focused design thinking.
  • An alternative approach is to embed or align security engineers directly with product development teams. This can be more effective in organizations where reporting structures and accountability are key considerations.
  • Some organizations draw Security Guardians from all job types. The program can also be used to focus on uplifting developers and broad security culture.
  • You must regularly inspect the outcomes delivered by the Security Guardians program and use the information to make incremental improvements as the program matures.

For additional support building a Security Guardians program, contact your AWS account representative and they will get you in touch with a specialist who can help you develop your program.

 
If you have feedback about this post, submit comments in the Comments section below. If you have questions about this post, contact AWS Support.
 

Mitch Beaumont Mitch Beaumont
Mitch is a Principal Solutions Architect for Amazon Web Services based in Sydney, Australia. Mitch works with some of Australia’s largest financial services customers, helping them to continually raise the security bar for the products and features that they build and ship. Outside of work, Mitch enjoys spending time with his family, photography, and surfing.
Ana Malhotra Ana Malhotra
Ana previously worked as a Security Specialist Solutions Architect and was the Healthcare and Life Sciences (HCLS) Security Lead for AWS Industry, based in Seattle, Washington. As a former AWS Application Security Engineer, during her time with AWS Industry, Ana loved talking all things AppSec, including people, process, and technology. In her free time, she enjoys tapping into her creative side with music and dance.