[Paper Review] Serverless Computing: Current Trends and Open Problems

Ioana Baldini, Paul Castro+9 more|arXiv (Cornell University)|Jun 10, 2017

IoT and Edge/Fog ComputingComputer Science2 references64 citations

TL;DR

This paper surveys serverless platforms, architectures, and use cases, and outlines key challenges and open research problems to advance the field.

ABSTRACT

Serverless computing has emerged as a new compelling paradigm for the deployment of applications and services. It represents an evolution of cloud programming models, abstractions, and platforms, and is a testament to the maturity and wide adoption of cloud technologies. In this chapter, we survey existing serverless platforms from industry, academia, and open source projects, identify key characteristics and use cases, and describe technical challenges and open problems.

Motivation & Objective

• Survey existing serverless platforms across industry, academia, and open source.
• Characterize common architectural features and programming models of serverless systems.
• Discuss benefits, drawbacks, and ecosystem dependencies of serverless computing.
• Identify current challenges and open research problems to guide future work.

Proposed method

• Review and compare serverless platforms using defined architectural dimensions.
• Describe the event-driven, function-centric programming model.
• Discuss deployment models, scaling, cost, and security considerations across platforms.
• Analyze use cases and workloads suitable for serverless (event processing, API composition, API aggregation).
• Summarize known benefits and tradeoffs for consumers and providers.

Experimental results

Research questions

• RQ1 What are the boundaries and scope of serverless computing relative to FaaS, PaaS, MBaaS, and SaaS?
• RQ2 What are the main architectural and operational challenges facing serverless platforms today?
• RQ3 How do current serverless platforms compare in terms of cost, performance, programming model, and composability?
• RQ4 What open research problems must be addressed to broaden adoption and interoperability of serverless systems?
• RQ5 How can legacy code and stateful workflows be effectively integrated with serverless architectures?

Key findings

• Serverless platforms are primarily event-driven, stateless function executors that scale transparently and can scale to zero.
• Major platforms include AWS Lambda, Google Cloud Functions, Microsoft Azure Functions, and IBM OpenWhisk, plus open-source efforts like OpenLambda.
• There are consistent tradeoffs in cost, performance, resource limits, and ecosystem lock-in that influence platform choice and application design.
• An ecosystem of supporting services (state, logs, authentication, API gateways) is critical but increases dependency on provider ecosystems.
• Open problems span boundaries between service models, state management, long-running tasks, debugging, and cross-cloud interoperability.