Cosmos: A Cost Model for Serverless Workflows in the 3D Compute Continuum

TL;DR

Cosmos introduces a cost and performance-cost tradeoff model for serverless workflows deployed across the 3D Edge-Cloud-Space Continuum, decomposing costs into Invocation, Compute, State Management, Data Transfer, and BaaS. The framework provides closed-form cost components per function and layer, plus a Pareto-front optimization to balance cost and latency under budget and SLO constraints. A case study across AWS and GCP demonstrates that data transfer and state management dominate IO-intensive workloads (up to 75% of AWS costs and 52% of GCP costs), while BaaS costs dominate compute-intensive tasks (up to 83% AWS and 97% GCP). The work enables cross-layer, workload-aware optimization and suggests an intelligent framework to predict function-level costs and dynamically select processing layers under varying workload characteristics and SLOs, with practical implications for deploying efficient serverless workflows in heterogeneous environments.

Abstract

Due to the high scalability, infrastructure management, and pay-per-use pricing model, serverless computing has been adopted in a wide range of applications such as real-time data processing, IoT, and AI-related workflows. However, deploying serverless functions across dynamic and heterogeneous environments such as the 3D (Edge-Cloud-Space) Continuum introduces additional complexity. Each layer of the 3D Continuum shows different performance capabilities and costs according to workload characteristics. Cloud services alone often show significant differences in performance and pricing for similar functions, further complicating cost management. Additionally, serverless workflows consist of functions with diverse characteristics, requiring a granular understanding of performance and cost trade-offs across different infrastructure layers to be able to address them individually. In this paper, we present Cosmos, a cost- and a performance-cost-tradeoff model for serverless workflows that identifies key factors that affect cost changes across different workloads and cloud providers. We present a case study analyzing the main drivers that influence the costs of serverless workflows. We demonstrate how to classify the costs of serverless workflows in leading cloud providers AWS and GCP. Our results show that for data-intensive functions, data transfer and state management costs contribute to up to 75% of the costs in AWS and 52% in GCP. For compute-intensive functions such as AI inference, the cost results show that BaaS services are the largest cost driver, reaching up to 83% in AWS and 97% in GCP.

Figures (7)

• Figure 1: Deforestation detection scenario with on-ground and in-orbit processing with serverless for the 3D Continuum.
• Figure 2: Simplified serverless workflow for deforestation with main cost drivers along and workload characteristics of each function.
• Figure 3: Serverless workflow costs drivers, highlighting key cost drivers: Invocation, Compute, Data Transfer, and State Management (partial view).
• Figure 4: Cosmos Performance Cost Tradeoff Model highlighting the optimal line between latency and costs, with theoretical lowest cost and latency, and SLO constraints of 50 USD and 75ms (pointed line).
• Figure 5: End-to-end latency for serverless workflows for AWS(x86) and GCP