PaaS
Definition
Platform as a Service - cloud platform that provides development tools and infrastructure so you can focus on building applications.
Use Cases
- The New York Times: Processing and transforming large volumes of images for online publishing — Used a PaaS-style approach on AWS by deploying image-processing code to a managed platform (commonly cited as AWS Elastic Beanstalk) and integrating it with other AWS services for storage and job handling (Faster processing and easier scaling for peak publishing demand, with reduced operational overhead compared to managing servers directly)
- Stack Overflow: Hosting and scaling a high-traffic Q&A web application — Used Microsoft Azure App Service to run web workloads with managed deployment, scaling, and platform maintenance handled by Azure (Simplified operations and improved ability to scale application capacity without managing underlying operating systems)
- Snapchat: Running backend services with rapid iteration and scaling needs — Used Google App Engine for parts of its backend to take advantage of managed deployment and autoscaling characteristics of PaaS (Accelerated development and scaling while offloading server management to the platform)
Provider Equivalents
- AWS: AWS Elastic Beanstalk
- Azure: Azure App Service
- GCP: Google App Engine
- OCI: Oracle Cloud Infrastructure (OCI) Application Platform (APaaS) / Oracle Visual Builder
Frequently Asked Questions
- What's the difference between PaaS and IaaS?
- IaaS gives you raw infrastructure like virtual machines, networks, and disks—you manage the operating system, runtime, patches, and scaling. PaaS sits higher: you deploy your code and the provider manages the servers, OS, runtime, and often scaling, so you can focus more on building the application.
- When should I use PaaS?
- Use PaaS when you want to ship applications quickly, avoid managing servers and OS patching, and benefit from built-in deployment and scaling. It’s a good fit for web apps, APIs, and background workers where standard runtimes (Node.js, Java, .NET, Python, etc.) work well and you don’t need deep control of the underlying OS.
- How much does PaaS cost?
- Pricing usually depends on the compute size/instance type, number of instances, runtime hours, and add-ons like managed databases, storage, logging, and outbound bandwidth. Many platforms offer free tiers or low-cost dev plans, but production costs rise with traffic, autoscaling, and premium features (private networking, higher availability, dedicated instances).
Category: cloud
Difficulty: intermediate
Related Terms
See Also