The field of cloud computing is witnessing significant advancements in autoscaling and serverless computing. Researchers are exploring innovative approaches to improve the efficiency and scalability of cloud resources, particularly in the context of dynamic and unpredictable workloads. One of the key trends is the adoption of reinforcement learning and artificial intelligence to optimize resource allocation and autoscaling. This involves developing adaptive algorithms that can learn from stream characteristics and make decisions in real-time. Another area of focus is the development of serverless computing frameworks that can efficiently process big data queries and provide low-latency responses. Additionally, researchers are working on designing benchmark suites for spatiotemporal database systems to better understand their quality of service behavior. Noteworthy papers in this area include the Archetype-Aware Predictive Autoscaling system, which reduces SLO violations by up to 50% and improves response time by 40%. The Multi-Agent Reinforcement Learning-based In-place Scaling Engine is another notable work, which enables seamless and dynamic resource scaling in edge-cloud systems. The KIS-S framework, which combines a GPU-aware Kubernetes Inference Simulator with a Proximal Policy Optimization-based autoscaler, also shows promise in improving the efficiency and scalability of GPU-accelerated environments.
Advancements in Autoscaling and Serverless Computing
Sources
Archetype-Aware Predictive Autoscaling with Uncertainty Quantification for Serverless Workloads on Kubernetes
Designing Adaptive Algorithms Based on Reinforcement Learning for Dynamic Optimization of Sliding Window Size in Multi-Dimensional Data Streams