The paper addresses the challenge of resource allocation for Industry 4.0 applications in remote industrial sites, particularly those with limited or no cloud access. The authors propose a serverless fog federation approach to dynamically federate nearby fog systems, enabling cloud-like elasticity and seamless resource allocation. They develop a probabilistic method, ProPar, to partition micro-service workflows across the federation and a resource allocation method, Maximum Probability (MR), to assign partitions to fog systems based on their likelihood of meeting deadlines. The experimental results show that ProPar can improve the deadline meeting rate by up to 35% compared to existing methods, demonstrating the effectiveness of their proposed solutions in handling demand surges and ensuring robust, latency-aware processing of Industry 4.0 applications.The paper addresses the challenge of resource allocation for Industry 4.0 applications in remote industrial sites, particularly those with limited or no cloud access. The authors propose a serverless fog federation approach to dynamically federate nearby fog systems, enabling cloud-like elasticity and seamless resource allocation. They develop a probabilistic method, ProPar, to partition micro-service workflows across the federation and a resource allocation method, Maximum Probability (MR), to assign partitions to fog systems based on their likelihood of meeting deadlines. The experimental results show that ProPar can improve the deadline meeting rate by up to 35% compared to existing methods, demonstrating the effectiveness of their proposed solutions in handling demand surges and ensuring robust, latency-aware processing of Industry 4.0 applications.