Deadline Sensitive And Function Placement In Multi-tier Serverless Platform

Serverless architecture has emerged as a distributed computing paradigm that raises the question of how to efficiently place functions in a distributed environment to meet deadlines and optimize resource utilization. This paradigm, also known as Function-as-a-Service (FaaS), provides businesses with the advantage of relinquishing resource and infrastructure management, allowing them to focus on their core business logic. Infrastructure and resource provisioning becomes the responsibility of cloud service providers. This proposal aims to address the challenge of function placement in a hierarchical or distributed edge-cloud environment, specifically for time-sensitive tasks where meeting deadlines is critical, considering the heterogeneity of compute nodes in terms of CPU, RAM, and memory. Our approach involves exploring the optimal strategy for function placement by utilizing offline data collection techniques and employing a non linear machine learning model for the placement strategy. To investigate this problem, we focus on deadline-sensitive Function within the context of a video surveillance camera application. We examine their placement across heterogeneous edge-cloud nodes, taking into account factors such as network latency and computation latency. The scope of our study is primarily centered around the deadline of tasks, the hierarchical nature of the edge-cloud infrastructure with geographically distributed nodes, the placement of functions, and the heterogeneity of serverless nodes. To simulate the execution environment, we employ Docker containers. By exploring an effective function placement strategy in a distributed edge-cloud environment, this research contributes to the optimization of resource utilization and timely execution of time-sensitive tasks within a serverless architecture.