Saarthi: An End-to-End Intelligent Platform for Optimising Distributed Serverless Workloads

Siddharth Agarwal, Maria A. Rodriguez, and Rajkumar Buyya · 2025

Read Saarthi: An End-to-End Intelligent Platform for Optimising Distributed Serverless Workloads by Siddharth Agarwal, Maria A. Rodriguez, and Rajkumar Buyya (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-06599.txt.

Description

Siddharth Agarwal, Maria Rodriguez, and Rajkumar Buyya present Saarthi, a platform that tunes serverless cloud functions to each incoming request. Function-as-a-service computing bills only for execution, yet cold starts, fixed configurations, and workload-blind scheduling cause uneven performance and surprise costs. Static sizing wastes resources on small requests and starves large ones.

Saarthi reads the request payload to anticipate resource needs, sizes functions accordingly, and orchestrates requests across configurations. A redundancy mechanism guards against faults, while an integer linear program keeps the number of running functions near optimal for throughput and cost. Built on OpenFaaS, it raised throughput, cut costs, and held service targets with small added delay.

Catalog record

CREATORSiddharth Agarwal, Maria A. Rodriguez, and Rajkumar Buyya
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE56 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-06599

Browse the full library · About the Tower of Babel