[Paper Review] Understanding Cost Dynamics of Serverless Computing: An Empirical Study

Muhammad Hamza, Muhammad Azeem Akbar+1 more|arXiv (Cornell University)|Nov 22, 2023

Cloud Computing and Resource Management5 citations

TL;DR

This empirical study investigates the cost dynamics of serverless computing through interviews with 15 industry experts from 8 organizations. It reveals that while serverless is ideal for unpredictable workloads, it may not be cost-effective for high-scale applications, and introduces a taxonomy for comparing serverless versus traditional cloud costs to guide optimization decisions.

ABSTRACT

The advent of serverless computing has revolutionized the landscape of cloud computing, offering a new paradigm that enables developers to focus solely on their applications rather than managing and provisioning the underlying infrastructure. These applications involve integrating individual functions into a cohesive workflow for complex tasks. The pay-per-use model and nontransparent reporting by cloud providers make it difficult to estimate serverless costs, imped-ing informed business decisions. Existing research studies on serverless compu-ting focus on performance optimization and state management, both from empir-ical and technical perspectives. However, the state-of-the-art shows a lack of em-pirical investigations on the understanding of the cost dynamics of serverless computing over traditional cloud computing. Therefore, this study delves into how organizations anticipate the costs of adopting serverless. It also aims to com-prehend workload suitability and identify best practices for cost optimization of serverless applications. To this end, we conducted a qualitative (interviews) study with 15 experts from 8 companies involved in the migration and development of serverless systems. The findings revealed that, while serverless computing is highly suitable for unpredictable workloads, it may not be cost-effective for cer-tain high-scale applications. The study also introduces a taxonomy for comparing the cost of adopting serverless versus traditional cloud.

Motivation & Objective

• To understand how organizations anticipate and manage costs when adopting serverless computing.
• To identify workload characteristics that make serverless cost-effective or inefficient.
• To develop a taxonomy for comparing serverless and traditional cloud cost models.
• To uncover best practices for optimizing serverless application costs based on real-world industry experience.
• To address the lack of empirical research on cost dynamics in serverless computing compared to traditional cloud models.

Proposed method

• Conducted semi-structured interviews with 15 experts from 8 companies experienced in serverless development and migration.
• Collected qualitative insights on cost estimation practices, workload patterns, and cost optimization strategies.
• Analyzed interview data using thematic analysis to identify recurring patterns and insights on cost behavior.
• Developed a comparative taxonomy for serverless versus traditional cloud cost models based on empirical findings.
• Focused on real-world organizational experiences to ground cost dynamics in practical deployment scenarios.

Experimental results

Research questions

• RQ1 How do organizations estimate and manage costs when adopting serverless computing?
• RQ2 What types of workloads are most suitable for serverless computing in terms of cost efficiency?
• RQ3 What are the key challenges in predicting and controlling serverless costs due to opaque billing models?
• RQ4 What best practices do organizations use to optimize serverless application costs?
• RQ5 How do serverless cost dynamics compare to those of traditional cloud computing in real-world deployments?

Key findings

• Serverless computing is highly suitable for unpredictable or sporadic workloads due to its pay-per-use model.
• For high-scale, consistently running workloads, traditional cloud solutions may be more cost-effective than serverless.
• Cloud providers' opaque billing models significantly hinder accurate cost estimation and planning.
• Organizations often lack standardized methods for forecasting serverless costs, leading to budget overruns.
• A taxonomy for comparing serverless and traditional cloud costs was developed based on expert insights, highlighting key differentiators in cost behavior.
• Best practices for cost optimization include workload profiling, function size minimization, and monitoring cold start frequency.