Acceleration-as-a-{\mu}Service: A Cloud-native Monte-Carlo Option Pricing Engine on CPUs, GPUs and Disaggregated FPGAs

TL;DR

This paper introduces CloudiFi, a framework for deploying hardware accelerators as cloud services, demonstrating significant performance improvements in financial workload processing.

Contribution

It presents a novel framework that enables flexible deployment and comparison of accelerators in cloud microservices, addressing previous limitations.

Findings

Up to 485x response time improvement in microservices

Effective deployment of accelerators as cloud services

Enhanced workload performance in financial applications

Abstract

The evolution of cloud applications into loosely-coupled microservices opens new opportunities for hardware accelerators to improve workload performance. Existing accelerator techniques for cloud sacrifice the consolidation benefits of microservices. This paper presents CloudiFi, a framework to deploy and compare accelerators as a cloud service. We evaluate our framework in the context of a financial workload and present early results indicating up to 485x gains in microservice response time.