FedLess: Secure and Scalable Federated Learning Using Serverless Computing

Abstract

The traditional cloud-centric approach for Deep Learning (DL) requires training data to be collected and processed at a central server which is often challenging in privacysensitive domains like healthcare. Towards this, a new learning paradigm called Federated Learning (FL) has been proposed that brings the potential of DL to these domains while addressing privacy and data ownership issues. FL enables remote clients to learn a shared ML model while keeping the data local. However, conventional FL systems face several challenges such as scalability, complex infrastructure management, and wasted compute and incurred costs due to idle clients. These challenges of FL systems closely align with the core problems that serverless computing and Function-as-a-Service (FaaS) platforms aim to solve. These include rapid scalability, no infrastructure management, automatic scaling to zero for idle clients, and a payper-use billing model. To this end, we present a novel system and framework for serverless FL, called FedLess. Our system supports multiple commercial and self-hosted FaaS providers and can be deployed in the cloud, on-premise in institutional data centers, and on edge devices. To the best of our knowledge, we are the first to enable FL across a large fabric of heterogeneous FaaS providers while providing important features like security and Differential Privacy. We demonstrate with comprehensive experiments that the successful training of DNNs for different tasks across up to 200 client functions and more is easily possible using our system. Furthermore, we demonstrate the practical viability of our methodology by comparing it against a traditional FL system and show that it can be cheaper and more resourceefficient.

Serverless on Machine Learning: A Systematic Mapping Study

“…The Federated Learning-based architecture was proposed in the primary dataset [47], [56], [66]. This computing model supports edge computing, where the processing edges can learn from a shared machine learning model while keeping the model training on remote clients, followed by global aggregation of the updated model parameters.…”

“…This keeps the training data local, which provides privacy and security benefits. Grafberger et al [56] considers that challenges of FL systems such as scalability, complex infrastructure management, and wasted computing can be solved with the Function-as-a-Service (FaaS) paradigm. However, it is necessary to be aware of the threats caused by malicious participants.…”

“…Several additional security measures can be applied, where only authorized and authenticated entities can invoke client functions. A practice of security between clients was applied in [56], where the FL server allows clients authenticated to read only from a shared global model and write back results without access to other clients. Another security measure was applied in [56], where HTTP function requests exchanges can be encrypted using Transport Layer Security (TLS).…”

MetaFaaS

“…A few efforts [30] in the literature focus on using the FaaS architecture to provide a costefficient, resource provisioning framework, enabling predictable performance for ML/DL workloads. A recent work [11], builds a framework for implementing federated learning using FaaS, which is slower albeit cheaper and more resource efficient in the long run. In spite of meta-learning architectures being compute-intensive and having relatively large training times, we have not come across any work in the literature, that employs FaaS to accelerate the training or re-training process of meta-learning architectures.…”