horizontal scaling
Handling more load by running more copies of a process, often on more machines, rather than making one machine bigger.
In more detail
Making one machine bigger is vertical scaling, and it has a ceiling. Horizontal scaling needs stateless, share-nothing processes, with something such as a load balancer spreading requests, so any copy can serve any request. Scaling back in safely also means draining work off a copy before it disappears.
Related terms
- vertical scaling — Handling more load by giving one machine more processors or memory, instead of adding more machines as horizontal scaling does.
- stateless — Describes a component or service that keeps no data from one request to the next, so each request carries everything needed to handle it.
- load balancer — A server that spreads incoming requests across several copies of a service so no single copy is overloaded.
- Auto Scaling group — An EC2 feature that keeps a desired number of identical instances running, replacing failed ones and adding or removing them as load changes.
- stateless process — A process that keeps nothing it needs between requests in its own memory or disk, so it can be restarted or replaced at any moment.
- process formation — The set of process types an app runs and how many of each are running.
- concurrency — Having several pieces of work in progress during overlapping periods of time, such as requests being handled at once.
- replica — One of several identical copies of a service or database running at the same time.
- stateless — Keeping no data between requests, so any copy of the service can handle any request.
- stateful — Holding data that must be kept and copied, such as a database or a shard.
Taught in
- Elasticity and Scale — Cloud Infrastructure & Hyperscaler Architecture
- Forty Terabytes Nobody Opens: Storage And Databases — CompTIA Cloud+ (CV0-004) Exam Prep
- Nine In The Morning, Every Monday: Scaling — CompTIA Cloud+ (CV0-004) Exam Prep
- Sixty Six Of Eighty Five: The Mock And Its Report — CompTIA Cloud+ (CV0-004) Exam Prep
- Build, Release, Run And Processes — DevOps & Site Reliability Engineering
- Port Binding And Concurrency — DevOps & Site Reliability Engineering
- Admin Processes And The Open Source Update — DevOps & Site Reliability Engineering
- Requirements: The Constraints That Shape The System — Scalable System Design & Distributed Architecture
- Capacity Planning As A Design Input — Scalable System Design & Distributed Architecture
References
- AWS Well-Architected Reliability pillar: use highly available network connectivity and scale horizontally
- The Twelve-Factor App: VIII. Concurrency
- Kubernetes documentation: horizontal pod autoscaling
- Azure Architecture Center: Design to scale out