What is API security?

As a field of web application security, API security is crucial for protecting modern applications that commonly rely on web services communicating via APIs to exchange data with users and other systems. For microservice architectures, entire apps are built using loosely coupled services that rely on API calls for external and also internal communication.

As developers are pushed to build applications and APIs more and more quickly, the use of AI coding assistants is on the rise. Looking at this from a security lens, that means more and more vulnerabilities are introduced into API code as these AI tools lack the necessary security awareness to support security coding practices effectively.

Compared to user interfaces, APIs provide a less visible way for attackers to access application data, including potentially sensitive information. This makes APIs a prime target and a significant source of data breaches that lead to business and personal data exposure. When you factor in the millions of IoT (Internet of Things) devices worldwide that rely on web APIs, successful API attacks can even allow malicious hackers to compromise some physical security measures or use internet-facing devices as entry points into internal systems.

REST APIs are by far the most common API type, used by over 85% of organizations that work with APIs according to a 2023 report. REST (REpresentational State Transfer) is not a strict protocol but an architectural style for building web applications and services, with JSON being the typical data interchange format.

In contrast, SOAP (Simple Object Access Protocol) is an XML-based API type where requests have to conform to a predefined schema. While less common than REST and slowly declining in popularity compared to GraphQL, SOAP APIs are still used in business applications.

A relative newcomer compared to REST and SOAP, GraphQL is a data query and manipulation language for building database access APIs that is rapidly gaining popularity, with up to 30% of organizations reporting they have some GraphQL APIs as of 2023.

Also worth mentioning is gRPC—a specialized API format designed specifically for high-performance microservice-based applications but also gaining popularity for mobile application backends. According to the same report, just over 10% of API developers were building with gRPC in 2023.

An application programming interface only serves as an intermediate layer for accessing an underlying application or system. This means you always have to think about API vulnerabilities on two levels:

• Vulnerabilities in the API itself: Only authorized and valid API requests should be passed on to the application. Compromising API safeguards allows attackers to break or bypass authorization, gain access to an API, and send malicious requests to the app. API vulnerabilities include weak or unprotected API keys, broken authentication mechanisms, failure to enforce end-to-end API traffic encryption with SSL/TLS (Transport Layer Security), and susceptibility to DDoS (Distributed Denial of Service) attacks through inadequate rate limiting.
• Vulnerabilities in the underlying application: To an attacker, an API endpoint is merely an extra application surface to probe and attack. Once API-level protections are broken or bypassed, malicious actors can target many common security vulnerabilities through API calls to attempt injection attacks such as SQL injection, command injection, and cross-site scripting (XSS). Server-side request forgery (SSRF) vulnerabilities are especially dangerous in the context of APIs as they can expose access to backend systems that weren’t supposed to be public-facing.

Vulnerable APIs can add a wide variety of security risks to your overall cybersecurity picture. The API Security Top 10 maintained by OWASP (the Open Web Application Security Project) is a popular resource that lists the most common API risk categories but focuses mostly on secure API design. A broader approach is to think of API security risks as being related to:

• Authorizing and authenticating access: Authorization failures can occur on the level of objects (broken object-level authorization, aka BOLA), object properties, and app functions (broken function-level authorization). Broken authentication is another major risk category, with unauthenticated API endpoints being a common data breach vector for sensitive data exposure.
• Limiting access: All API access must be constrained and managed to mitigate security threats such as server resource exhaustion, mass data extraction, and other attempts to abuse API functionality, for example through brute-force enumeration.
• Inventory management: Running unmaintained old versions of endpoints or entire APIs provides threat actors with an easy starting point, greatly increasing the risk of unauthorized access. Ideally, organizations should know and document all their APIs and endpoints, both private and public, though this is rarely achieved in practice.
• Configuration: Security misconfigurations, most notably misconfigured security headers, are a common source of risk for web applications and APIs alike, introducing security risks that are beyond developer control.
• Security vulnerabilities: API attacks are often only one small part of a larger application attack targeting specific security flaws. SSRF vulnerabilities can be especially impactful, allowing attackers to manipulate URLs to get access to remote resources via APIs.