Energy-Efficient Deployment of Stateful FaaS Vertical Applications on Edge Data Networks

Overview

• 5G and beyond networks enable the deployment of vertical applications, which can be enhanced by network slicing and edge computing to create isolated environments for executing customer services.
• While serverless computing (also known as Function as a Service or FaaS) has gained popularity in cloud-native environments, its adoption at the edge is lagging, particularly due to the need to support stateful tasks that are common in applications like cognitive services.
• This research explores the emerging paradigm of stateful FaaS with lightweight task abstractions using WebAssembly, a technology that enables efficient execution of code on edge devices.

Plain English Explanation

The paper examines how the latest 5G and future wireless technologies can be used to deploy specialized applications for different industries (known as "vertical applications"). These applications can benefit from network slicing and edge computing, which create isolated environments for running customer services closer to the devices that use them.

One promising technology for running these services is serverless computing, also called "Function as a Service" (FaaS). FaaS allows developers to run small pieces of code (called "functions") without needing to manage the underlying infrastructure. However, adopting FaaS at the edge has been challenging, especially for applications that need to maintain state (keep track of information) between function calls, like cognitive services.

The researchers in this paper study a new approach called "stateful FaaS" that uses a lightweight technology called WebAssembly to enable stateful tasks to run efficiently on edge devices. They assess the tradeoffs between stateless and stateful approaches in terms of energy consumption, which is a critical factor for edge computing.

Technical Explanation

The paper explores the use of stateful FaaS with lightweight task abstractions in WebAssembly as a way to deploy interdependent, stateful tasks on edge computing resources. The researchers compare the energy consumption of stateless vs. stateful approaches, considering both the power used for processing and communication.

To do this, they derive a mathematical model to estimate the energy consumption of a given workload based on its characteristics. They then use this model in extensive simulations to determine the impact of key factors and assess the energy trade-offs between stateless and stateful approaches.

The results provide insights into the conditions under which stateful FaaS can be more energy-efficient than stateless FaaS for edge computing, particularly in scenarios with energy-constrained IoT applications that require maintaining state across function invocations.

Critical Analysis

The paper presents a thorough analysis of the energy implications of stateful FaaS on edge computing, which is an important consideration for empowering IoT applications in real-world settings. However, the research is limited to a simulation-based approach and does not include any experimental validation on actual edge devices.

Additionally, the paper acknowledges that the proposed mathematical model may not capture all the nuances of real-world edge computing environments, such as the impact of network conditions, hardware heterogeneity, and dynamic workload changes. Further research and empirical evaluation would be necessary to fully understand the practical implications of stateful FaaS at the edge.

Conclusion

This research explores the potential benefits of using stateful FaaS with WebAssembly to deploy interdependent, stateful tasks on edge computing resources. The paper provides a detailed analysis of the energy trade-offs between stateless and stateful approaches, suggesting that stateful FaaS can be more energy-efficient in certain scenarios involving energy-constrained IoT applications. These findings have important implications for the future development and adoption of edge computing technologies to support a wide range of vertical applications.