Memory orchestration mechanisms in serverless computing: a taxonomy, review and future directions.

Serverless computing has become very popular in recent years due to its flexibility and cost efficiency. Serverless computing is a cloud computing model that allows developers to write and deploy code without managing the underlying infrastructure. Serverless computing has been welcomed due to its scalability, affordability, and ease of use. However, memory orchestration in serverless computing can be challenging due to the ephemeral nature of serverless functions, which are short-lived and stateless. This article has reviewed memory orchestration mechanisms and then presented a classification of memory orchestration mechanisms in serverless computing, which are classified into three main methods: Machine learning-based approach, Heuristic-based approach, and framework-based approach. The advantages and disadvantages of each mechanism as well as the challenges and performance metrics affecting their effectiveness have been investigated. Each memory orchestration approach, whether heuristic-based, framework-based, or machine learning-based, has its own advantages and disadvantages. Different mechanisms are suitable for different use cases. Therefore, it is important to carefully evaluate the trade-offs between performance, cost, and complexity when choosing a memory orchestration mechanism. Finally, the paper identifies several future research directions for memory orchestration in serverless computing, which include developing memory orchestration mechanisms and dynamic memory management, integrating memory orchestration mechanisms with other resource management mechanisms, security, memory sharing, compression, pricing, and investigating the relationships between memory orchestration with cost and performance in serverless computing. [ABSTRACT FROM AUTHOR]