🤖 AI Summary
To address high deployment costs and performance instability caused by frequent interruptions of preemptible (Spot) instances in public cloud–based microservices, this paper proposes a Spot-aware elastic scheduling framework. The framework integrates a genetic algorithm with node-termination fault tolerance to enable application workload–aware, fine-grained resource recommendation and zero-perception interruption recovery. It further combines adaptive cluster autoscaling (CA) with containerized scheduling to build an end-to-end elastic system on Kubernetes. Evaluated in a real public cloud environment, the framework reduces average deployment cost by 38.7% compared to baseline strategies, achieves service availability exceeding 99.95%, and ensures autoscaling response latency under 12 seconds. Its core innovation lies in the first deep coupling of evolutionary optimization with Spot-instance fault tolerance—jointly optimizing cost efficiency and service resilience.
📝 Abstract
Microservices architecture, known for its agility and efficiency, is an ideal framework for cloud-based soft-ware development and deployment, particularly when inte-grated with containerization and orchestration systems that streamline resource management. However, the cost of cloud computing remains a critical concern for many organizations, prompting the need for effective strategies to minimize expenses without compromising performance. While cloud platforms like AWS offer transient pricing options, such as Spot Pricing, to reduce operational costs, the unpredictable demand and abrupt termination of spot VMs introduce considerable challenges. By leveraging containerization alongside intelligent orchestration, it is possible to optimize micro services deploy-ment costs while maintaining performance requirements. We present SpotKube, an open-source, Kubernetes-based so-lution that employs a genetic algorithm for cost optimization. Designed to dynamically scale clusters for micro service applications on public clouds using spot pricing, SpotKube analyzes application characteristics to recommend optimal resource al-locations, ensuring deployments remain cost -effective without sacrificing performance. Its elastic cluster autoscaler adapts to changing demands, gracefully managing node terminations to minimize disruptions in system availability. Evaluations conducted using real-world public cloud setups demonstrate SpotKube's superior performance and cost efficiency compared to alternative optimization strategies. GitHub: https://github.com/SvotKube/SvotKube