DevOps is an agile development approach that aims to reshape how organizations structure their software development practices. It involves integrating development (Dev) and operations (Ops) into a single, continuously reinforcing development pipeline, allowing the combined DevOps teams to simultaneously develop, update, and deploy software products.
While DevOps greatly improves the agility and efficiency of the software development lifecycle (SDLC), it can come at the expense of security, which is often neglected or relegated to an afterthought. According to a SANS survey , only 46% of developers address security concerns in the early stages of the development pipeline, and only about half get round to fixing major vulnerabilities.
An example of how serious the security threat is to organizations that have adopted DevOps practices is Uber’s breach in 2016. This breach exposed the information of around 600,000 drivers and 57 million customers.
DevOps practices enable shorter development and release cycles and faster responses to customer demand. However, there are a number of issues affecting DevOps security.
1. Keeping Pace with DevOps
A major hurdle for security teams is the struggle to adjust to the speed of the DevOps process. In traditional waterfall development models, long development cycles allowed for more breathing space, and security testing would take place at the end of the SLDC, once the other aspects of software development had been completed. Developers could then go back and fix any vulnerabilities that were discovered.
In the DevOps model, however, release cycles that were once several months have been condensed to mere weeks or even days, and there is less room for delays in today’s consumer market. If security teams continue to work as they are used to, only performing security checks at intervals of several months, they cannot ensure the security of each iteration, exposing the product to an increased risk of exploitation by attackers.
2. Communication between Security and DevOps Teams
The goals of developers tend to differ from those of security teams. While developers aim to release software as soon as possible, security teams slow them down, demanding they fix all security vulnerabilities before deploying an application.
Furthermore, security teams often fail to effectively communicate with development teams, who may not adequately track or resolve security issues. This results in a phenomenon known as “silo-ization”, whereby security and development teams are separated and unfamiliar with each other’s roles. The lack of communication further slows down the development process because developers are unable to fix security issues as they build their software.
3. Workplace Culture
Often accompanying “silo-ization” is an atmosphere of resistance to change within the workplace. Developers and security personnel, as well as management or the business department, are often comfortable with familiar procedures and may find it difficult to incorporate security into the DevOps model.
Organizations, and especially developers, may prioritize quick release cycles, treating security as an afterthought. They perceive security checks as a drain on the agile development process, slowing it down. However, the technical debt accumulated by kicking the can down the road can prove to be far more costly and time-consuming later on.
4. Cloud Security Issues
DevOps environments typically take advantage of cloud computing and open-source tools for scalability and agility, but this raises additional concerns regarding cloud security. A small error or vulnerability in the cloud environment can easily spread and result in malfunctions or attacks. Security in a public or hybrid cloud differs from traditional data center security because it is harder to establish a secure perimeter, especially when you are using a shared network. You cannot rely solely on firewalls to protect your cloud environment.
5. Inadequate Access Controls and Unmanaged Secrets
DevOps tools typically utilize secrets, such as privileged account credentials and API tokens. If you lack strong privileged access controls, you may be exposing your system to attack. Weak access controls allow unauthorized outsiders or rogue insiders to penetrate your network and steal information, disrupt operations, or even highjack your infrastructure. Systems that use DevOps are especially vulnerable given the interconnected nature of the DevOps environment.
6. Inefficient Or Vulnerable Tools and Infrastructure
Most organizations run hybrid environments that use both cloud-based and legacy resources. While cloud-based and open source tools come with their own vulnerabilities, legacy resources are especially difficult to secure. Likewise, it may be challenging to integrate new tools or processes into your legacy infrastructure, affecting the flexibility and efficiency of your system. Over-reliance on legacy resources can make it more complicated to meet changing security and performance requirements.
To overcome the challenges facing organizations when implementing DevOps and agile workloads, you should consider taking the following steps:
1. Integrate Security into the SDLC
You can improve the security of your applications by expanding the DevOps philosophy to security and baking security into the software development lifecycle (SDLC). In a model sometimes referred to as DevSecOps, organizations shift security left, addressing security considerations early on in each development cycle. This allows the development team to work alongside your security team to reduce waiting time and technical debt.
2. Train Your Personnel
Security personnel may require new skills to keep up with the DevOps process. You should ensure that security practitioners can write code and work with APIs so they can automate some of their security tasks and collaborate with developers. Some organizations opt for retraining developers to perform security tasks. It is also important to educate staff about the importance of integrating security into the DevOps process.
3. Create a Culture of Cooperation
Perhaps the most challenging but nonetheless important step involves shifting your organization to a DevSecOps mindset. This means that everyone, including developers, shares responsibility for security, without leaving it entirely to security practitioners. The security team must work to bridge the communication gap and ensure that the rest of the organization understands security.
Developers must not sacrifice security in their pursuit of speed.
You should provide governance codes for your organization that incorporate security as a primary concern, and establish security protocols that will ensure that everyone knows how to address security.
4. Implement Cloud Security
In the cloud, security involves separating your infrastructure with logical access limitations, encryption, and network policies. Network segmentation helps minimize potential attack vectors by limiting visibility for an attacker. If you need access between separate trust zones, you can use a jump server secured with multiple layers of authentication. You can also take advantage of the open source community to stay on top of newly-discovered vulnerabilities. Public cloud providers tend to manage most security concerns for you, but you still need to ensure that you maintain visibility and monitor all access to your network.
5. Manage Your Credential Controls
Centralize and monitor all access to your system. You should limit access controls and privilege rights to essential users, removing credentials embedded into code, files, and service accounts. You should also ensure that passwords are kept separate from the code.
6. Automate Security Processes
You can save time and effort by automating security processes, especially repetitive ones like threat detection and code analysis. You cannot scale up to meet the demands of a large DevOps environment without the ability to automatically respond to threats. Automation is an area that is constantly expanding, and you can progressively automate more tools and process as you grow your organization. Automation also reduces risks related to human error and helps avoid downtime.
As security teams struggle to keep up with developers in the ever-accelerating race to deployment, vulnerabilities can be left ignored or undiscovered. To ensure that you can produce both high-quality and safe software within short release cycles, you must learn to fully integrate security into your DevOps process.
