Towards a Testbed for Scalable FaaS Platforms
🤖 AI Summary
Current FaaS platforms exhibit significant performance variability due to underlying architectural differences, yet there is a lack of customizable, research-oriented testbeds designed specifically for scalability evaluation. To address this gap, we propose a modular and configurable FaaS testbed framework that enables controlled, comparative experimentation across mainstream architectures—including Knative, OpenFaaS, and the AWS Lambda emulator. The framework integrates a benchmark workload generator, fine-grained performance monitoring, and end-to-end latency analysis tools to support quantitative modeling and empirical assessment of scalability behavior. We validate the testbed across diverse deployment scenarios: single-node, Kubernetes clusters, and edge environments—demonstrating both functional correctness and near-linear scalability. This infrastructure provides a reproducible, extensible foundation for systematic FaaS architecture design, evaluation, and optimization.
Application Category
📝 Abstract
Most cloud platforms have a Function-as-a-Service (FaaS) offering that enables users to easily write highly scalable applications. To better understand how the platform's architecture impacts its performance, we present a research-focused testbed that can be adapted to quickly evaluate the impact of different architectures and technologies on the characteristics of scalability-focused FaaS platforms.
Problem
Research questions and friction points this paper is trying to address.
Evaluating impact of architecture on FaaS platform performance
Developing adaptable testbed for scalable FaaS platforms
Assessing technologies for scalability in cloud platforms
Innovation
Methods, ideas, or system contributions that make the work stand out.
Research-focused testbed for FaaS platforms
Evaluates impact of different architectures
Assesses scalability-focused platform characteristics
🔎 Similar Papers
No similar papers found.
