by Blake Dobbs | May 3, 2018 | Automation and Orchestration
With our final part in the Security Automation series, Kudelski Security will take a look at what our clients are doing to take their manual playbooks to the next level, using automation. Before we take a look at playbooks, a quick review of the key factors from our previous articles to ensure automation success.
Keys to Successful Automation
- Understanding your organization’s problem that you are trying to solve
- Understanding your organization’s environment as it exists currently
- Understanding the maturity of your program and mapping it to a cybersecurity framework
- Identifying the business risk areas that automation can solve
- Identifying the common business issues across teams in the organization
- Designing and documenting manual processes to allow for future automation
With a firm grasp on these automation keys, let’s take a look at a common use case for most environments that are ripe for automation: Firewall Requests.
Firewall Requests
With most organizations protecting their network perimeter with a firewall (or ten), a question we get asked often is how to make that process more efficient, even if without automation, but preferably with automation. Let’s take a look at a fairly common workflow for generating firewall changes.
One of the first things we review is how the teams interact with one another for any workflow. In this example, each team is shown in a different color because they interface with each other using a different mechanism, whether that be by email, chat, or by separate IT Service Management (ITSM) tools (e.g. ServiceNow). While everyone in the technology world has been looking for the holy grail of a single pane of glass for all tools, for processes like these and most other IT operations processes, it is fairly routine to have all parties involved using the same means for communicating and documenting the tasks in a workflow. Using the same tool or mechanism for communication allows you to then create a separate channel or project for each workflow, and each channel or project includes only the team members needed for that workflow. For example, most network operation teams use an ITSM and have their own project for handling day to day activities, and the same for the firewall team, application team, etc. By leveraging that same ITSM and creating a new project that includes all teams that are required for that workflow, there is a central location for all communication and interaction, for both a manual and automated process.
Once a centralized location or mechanism for handling these requests has been agreed upon, it really facilitates open discussions and interactions on what the actual process should look like, where changes can be made, and where parts of the process are being duplicated. We have seen the most success when the teams get together, in the same room (crazy right?), and iron out the process to where all parties are satisfied and unified. From a manual process, it becomes a much cleaner simpler process that looks like the figure below.
From this cleaner process, we can apply automation to handle the interactions between the central location or mechanism for communication (i.e. ITSM, email) and the relevant technology to handle the firewall changes via automation platforms or custom API gateways. The final playbook comes out like the figure below.
From this playbook, we can see how everything comes together. The interactions between the ITSM and Firewall Management System are handled via web-hooks or API calls either after the push of a button from the ITSM workflow or via the custom-built automated workflow in the Firewall Management System (e.g. FireMon). This allows for maintaining a human interaction in the loop but drastically reduces the required work for that individual. That individual or team is also only using one system for this interaction, with all data flowing through the ITSM for process handling, thus reducing the number of systems required for access, and creating a much easier audit trail for long-term auditability and accountability.
Reviewing Keys to Automation Success
Let’s take a look at a few of the key areas for successful automation:
- Understanding your organization’s problem that you are trying to solve
- In our example, the organization’s problem was a decentralized firewall request process that led to many interactions with multiple systems to accomplish the same task.
- Understanding your organization’s environment as it exists currently
- In our example, the organization’s environment was a typical enterprise environment, with teams existing in different segments. These teams are required to interact with one another to accomplish certain tasks but do not see the inefficiencies in the overall process. Once able to understand each team’s process and responsibilities, the realization became clear that there were many areas they were able to consolidate.
- Identifying the common business issues across teams in the organization
- In our example, the common business issue was each team had their own process for handling the same task, albeit a smaller portion. With each team having to create and manage their own ITSM workflow and incidents and not able to interact with other teams, there were lots of tasks being repeated, and no centralized location for communication or data retention.
- Designing and documenting manual processes to allow for future automation
- In our example, the creation of the visual workflow for the legacy process was a key driver to have each team get together and spend quality time redesigning the process. The process was disjointed and repetitive for no real reason and a simpler, more consolidated, process was quickly realized and developed. By consolidating the process and having each team be a key stakeholder in the process, the benefits of automation could be realized by augmenting what the teams already handle on a daily basis.
Final Thoughts and Questions
Many companies are beginning to understand that automation will not soon replace human interaction in IT, but there are many processes that can be improved by it. Automation should not be used to fix a broken process, but to give information to those who can fix the problem in a faster, repetitive, and less expensive fashion. This will allow your superstar employees to continue to be superstars, freeing them from remedial tasks, and allow for less experienced employees to get all the information they need without having to needlessly search for the correct resource to engage for assistance. From a CISO’s perspective, it comes down to what problems am I having, and is there a sliver of a chance that automation can be leveraged to maximize the return on investment. At Kudelski Security, we are happy to assist with understanding what issues exist and where there are potentials for automation or just assistance helping to make processes more efficient, just let us know how we can help.
by Blake Dobbs | Dec 18, 2017 | Automation and Orchestration
In our previous Security Automation series post, we identified areas that should be reviewed to allow for the most success with automation. Those areas included identifying the problems, dealing with the environment, and looking for frameworks that can apply a solid foundation for the security program and its automation success. In this post, we will look at how to apply those ideas to start building a security program that is designed for automation to have a key role in the program’s success.
Identifying and Determining Risk
After reviewing the problems your organization is facing, and observing those problems in your environment, the next step is to identify areas of risk and quantifying those areas, to determine the risk level for each set of problems. How do you quantify these risk levels?
- Data gathered from monitoring tools
- Data gathered from business owners
- Internal or external scans
Data gathered from monitoring tools
Using monitoring tools such as SIEMS or log-aggregators allow for a centralized location for events happening in your environment. These tools provide vital information about the systems in your environment, and serves as the launch point for many processes and tasks, and with the vast amount of information in a centralized place, allows automation to have information readily available and accessible. Simple items, such as hostnames, users/groups, IP addresses, and application names are items that security analysts spend a lot of time gathering for each incident, and are required to get the same information for every event. Having the right monitoring tools along with automation can serve that information up to the analysts automatically for every new event.
Data gathered from business owners
Getting information from business owners, managers, or individual teams about how much risk is associated with each product, application, or service is crucial to the overall security picture for your organization. This information is often not represented well, due to teams not communicating with each, not getting the information on a scheduled basis, or not having a central place to store or visualize the information. This is where frameworks and services such as Kudelski Security’s Secure Blueprint really assist organizations determine the risk for each business owner, and rolling those risks up with the security posture of the organization.
Internal or external scans
Running continuous or scheduled scans of your environment for vulnerabilities, new or removed systems, and network changes allow you to get an understanding of what needs attention. These scans and tools give a technical picture of the potential holes in your environment, and allow both the business owner and your security team to determine the risk associated with each system, and which process to apply to remediate these systems. Automation can play a large role with these scans, from running the scans on a scheduled basis, to moving high risk systems to a quarantined zone, or to remediating the systems based on overall risk score automatically.
Identifying Common Business Issues
As areas of risk come into focus, taking a look at those areas to determine any commonality, identifying common challenges and platforms, even if it spans multiple teams. In large organizations, one of the largest challenges with both security and automation, is that each team will not communicate with each other well, often using similar platforms and duplicating the work. What common issues are occurring in each risk area?
- Fatigue from overall volume, leading to high mean time to response (MTTR)
- Lack of relevant information, leading to multiple teams responding to same issue
- Lack of documented processes
Fatigue from overall volume, leading to high mean time to response (MTTR)
Large majorities of security and IT teams across all verticals deal with alert fatigue, either from not having enough personnel to deal with the events, not properly configuring their security tools, or from not have a well-defined process for handling the events. These reasons lead to the teams not responding quickly enough, or in some cases not responding at all, with big events getting lost in the noise. It is important to recognize when these teams are having difficulty responding to events in a timely manner, especially when there are multiple teams that collaborate with each other.
Lack of relevant information, leading to multiple teams responding to same issue
Many security teams spend a large bulk of their time searching for relevant information to appropriately respond to an incident. Many times, multiple teams receive the same event, and work on the event in parallel with other teams, not knowing what the other team is doing. Building a security environment that has centralized reporting and monitoring, centralized case management, and a mandate from the executive level to communicate with other teams really allow the security team to thrive. Adding automated processes to those teams just puts more time back into the analysts workflow, augmenting their skills to respond to the event in a more efficient manner.
Lack of documented processes
All companies have a process for handling events, whether that is just having an analysts “fix” it, or a detailed workflow diagram that is followed religiously. Reliability is the biggest key for security processes, can they be completed over and over, the same way every time. When security teams do not have a reliable, documented process, it leads to having gaps in your ability to handle the events. If analysts handle events on their own, it is often found that those analysts spend more time for each event, leading to other events potentially falling through the cracks, or the same event not being handled the same way the next time. Another key for processes is they must be just flexible to change, but those changes need to be documented or version controlled as the business needs change.
Designing Processes for Automation
With a solid understanding of the risks, the common issues, and the frameworks that help build the case for automating a task to help better align with a business objective, the foundation is set to allow automation to thrive. Beginning an automation process without these key areas is automating for the wrong reasons, and generally leads to homegrown scripts and applications that require more maintenance than benefit. When starting to design an automated process, some key areas to have mapped out:
- What system(s) is the process targeting?
- What level of human interaction is wanted?
- Reporting success/failures of the process
What system(s) is the process targeting?
Knowing what system(s) to target with a process is vital to designing the process for automation. This allows for boundaries to be set for the automation to work within, keeping it from moving beyond its intended need. Knowing what system(s) also provides a better understanding of what information you will need to gather to interact with those system(s).
What level of human interaction is wanted?
Automation is not designed to replace your security team, only to augment them. Identifying key areas within the process for a human interaction to either approve the workflow to the next step, requiring someone to input a particular device target, or having someone audit the workflow once it has finished before pushing back into production.
Reporting success/failures of the process
After building simple or elaborate processes that automation can implement and really assist your security team, there has to be a way to measure the outcomes to map back into the overall security posture of the organization. By taking the automation journey this far, adding those measures back into the overall risk score for the organization allow for continuity in your security program and its risk posture.
With these areas mapped out, a documented workflow can be automated just by filling in the holes in the workflow with system info and/or adding an analyst approval step. These documented workflows that are being automated allow your team to spend less time getting information and more time responding to the incidents at hand. Don’t have a documented workflow? At Kudelski Security, we have built numerous custom workflows for customers after answering the above questions. Our team thrives on building effective processes for challenging but repetitive tasks, allowing your security team to focus on protecting your business.
Up Next
In the last part of this series, we look at taking security automation to the next level, improving playbooks, and bringing multiple assets into one workflow to improve overall security efficiency.
