Microservices

Even transient bottlenecks (very short resource saturations) can significantly degrade the performance of distributed systems due to their complex inter-service dependencies. Through detailed timeline analysis, we investigate system behavior to identify the root causes and cascading effects of these bottlenecks, providing insights into their mitigation and impact on overall system stability.

Loosely-coupled and lightweight microservices can create complex execution dependencies within the system. Cross-path dependencies arise when two execution paths partially share component microservices, potentially leading to runtime blocking effects. As a result, a slowdown in one execution path (e.g., due to a transient bottleneck) can indirectly delay other paths, causing widespread performance degradation.

Based the understanding of execution dependencies in microservices, we find that even a short resource saturation in a single microservice can trigger widespread performance degradation. To further explore these performance vulnerabilities, we introduce Grunt Attack—a novel low-volume DDoS attack that exploits execution dependencies to disrupt system stability.