Optimizing Serverless Computing

Abstract—Serverless computing abstracts infrastructure man- agement, hence provides scalability and cost-effectiveness. Per- formance bottlenecks related to this architecture include cold starts, resource inefficiency, and increased latency of execution. Strategies for optimizing serverless computing would involve con- tainer pooling and AI-based scheduling, to add to the resource- aware function execution and blockchain implementation ap- proaches. On an OpenFaaS and AWS Lambda platform, the above-proposed model realizes 40% lower cold-start latency and 25% less execution cost. The paper will further explore adaptive scaling-both for warming the cache and decentralized computation with load balancing added to further increase efficiency with multiple clouds. This research also introduces the possibility of blockchain-based serverless execution as an application that maintains security as well as transparency in computation processes. Index Terms—Serverless Computing, Cold Start Optimization, AI-Based Scheduling, Function-as-a-Service (FaaS), Image Pro- cessing, Load Balancing, Adaptive Scaling, Blockchain Integra- tion