DevSecOps Begins at the Application Layer
February 19, 2019

Brian Jenkins
StrongKey

In the fast-moving world of DevOps, security sometimes got left by the wayside on the way to the next iteration. But today's threat landscape is so perilous that developers need to have solid security top of mind as they design and build applications. This includes features like user authentication, digital signatures, and encryption.

Learning from Insecurity at Uber

Uber's operating procedures provide many cautionary tales, but this one gives a vivid description of why security is needed at every level and what can go wrong if it is not in place. An Uber software developer stored a service credential in application code to access sensitive information from the database. He then stored the code in a private repository in GitHub.

Including service credentials – a "shared secret" – inside software would, in itself, have been a violation of security best practices, since these can be compromised in many places besides the GitHub repository. This would include testing environments, staging machines and, of course, the production infrastructure itself.

Here's the rub: GitHub wanted software developers to protect their repositories from unauthorized access, so it had already implemented FIDO-based strong authentication. Despite the deployment of one of the strongest authentication protocols in the industry, GitHub neither encourages its users to sign up nor sign in with FIDO technology. Consequently, the Uber software developer used one or more shared secrets – username/password, one-time passcodes, etc. – to authenticate to the Uber repository.

But wait – there's more. Uber also automatically deployed its applications into Amazon Web Services (AWS) using yet another shared secret: an application programming interface (API) key with a secret key – in other words, a username and password.

The result was a perfect storm of insecure practices: using passwords to store software in a repository, containing passwords to access a protected database, using passwords to automatically deploy applications into the public cloud. It all eventually led to the compromise of sensitive data. Uber suffered a breach of 57 million passenger and driver records in 2016.

Security at the Application Level

This story included many security faux pas, but malicious actors only need to find one. FIDO's multiple uses make it an ideal security option. Though FIDO protocols were primarily designed to enable strong authentication to web applications, they can also support transaction authorization. Sadly, as noted above, organizations are apparently not using either feature in any consistent, meaningful way.

This leads to unnecessary losses. FIDO protocols not only have the potential to strengthen transaction security, but they eliminate the password nightmare that end-users go through while protecting them and web applications from many attacks on the internet.

Ransomware is a prime example of why organizations need application security. Such attacks work because applications allow authenticated users to modify files – encrypting them and deleting the original file – without secondary authentication and/or authorization. Consequently, malware executing on users' computers do so with full privileges of the user. FIDO digital signatures change that framework, leading to higher levels of security.

An additional security measure at this point is transaction-level authorization to not only deter transaction fraud but also to protect against ransomware. The protocols are available today; the tools are available now. All that is required is the resolve to implement these measures.

Laying a Secure Groundwork

One of the reasons bad actors continue their successful streak of data breaches is that organizations are still operating on the mistaken assumption that it is easier to deter "barbarians at the gate" than to protect sensitive data in the application. As a result, companies over-invest in network-based security tools – firewalls, anti-virus, malware detection, intrusion prevention, etc. – rather than also invest in the control mechanism that provides the highest level of data-protection: application-level encryption.

By adding application-level encryption from the ground up, software developers serve their users well. While some risk can be reduced by using FIDO-based strong authentication controls, reasonable data security requires multiple controls to deter attackers – a practice the security industry terms as "defense in depth."

The most secure data protection a company can hope to implement, without eliminating sensitive data from a system, involves encrypting and decrypting data within authorized applications (combined with a hardware-backed cryptographic key management system). Adding FIDO-based strong authentication creates a high level of risk mitigation. By building these security measures in from the start, DevSecOps professionals lay the groundwork for maximum-strength protection where it counts most.

Brian Jenkins is VP of Product at StrongKey
Share this

Industry News

March 06, 2025

Parasoft(link is external) is showcasing its latest product innovations at embedded world Exhibition, booth 4-318(link is external), including new GenAI integration with Microsoft Visual Studio Code (VS Code) to optimize test automation of safety-critical applications while reducing development time, cost, and risk.

March 06, 2025

JFrog announced general availability of its integration with NVIDIA NIM microservices, part of the NVIDIA AI Enterprise software platform.

March 06, 2025

CloudCasa by Catalogic announce an integration with SUSE® Rancher Prime via a new Rancher Prime Extension.

March 05, 2025

MacStadium(link is external) announced the extended availability of Orka(link is external) Cluster 3.2, establishing the market’s first enterprise-grade macOS virtualization solution available across multiple deployment options.

March 05, 2025

JFrog is partnering with Hugging Face, host of a repository of public machine learning (ML) models — the Hugging Face Hub — designed to achieve more robust security scans and analysis forevery ML model in their library.

March 05, 2025

Copado launched DevOps Automation Agent on Salesforce's AgentExchange, a global ecosystem marketplace powered by AppExchange for leading partners building new third-party agents and agent actions for Agentforce.

March 05, 2025

Harness completed its merger with Traceable, effective March 4, 2025.

March 04, 2025

JFrog released JFrog ML, an MLOps solution as part of the JFrog Platform designed to enable development teams, data scientists and ML engineers to quickly develop and deploy enterprise-ready AI applications at scale.

March 04, 2025

Progress announced the addition of Web Application Firewall (WAF) functionality to Progress® MOVEit® Cloud managed file transfer (MFT) solution.

March 04, 2025

Couchbase launched Couchbase Edge Server, an offline-first, lightweight database server and sync solution designed to provide low latency data access, consolidation, storage and processing for applications in resource-constrained edge environments.

March 04, 2025

Sonatype announced end-to-end AI Software Composition Analysis (AI SCA) capabilities that enable enterprises to harness the full potential of AI.

March 03, 2025

Aviatrix® announced the launch of the Aviatrix Kubernetes Firewall.

March 03, 2025

ScaleOps announced the general availability of their Pod Placement feature, a solution that helps companies manage Kubernetes infrastructure.

March 03, 2025

Cloudsmith raised a $23 million Series B funding round led by TCV, with participation from Insight Partners and existing investors.

February 27, 2025

IBM has completed its acquisition of HashiCorp, whose products automate and secure the infrastructure that underpins hybrid cloud applications and generative AI.