Compute Instance
Definition
Oracle's virtual machine service providing scalable compute resources, enabling businesses to run applications efficiently in the cloud.
Use Cases
- Zoom: Rapidly scaling video conferencing infrastructure during demand spikes — Used public-cloud virtual machine capacity (including AWS EC2) with autoscaling and load balancing to add compute quickly as meeting traffic increased. (Improved ability to handle sudden growth by adding capacity quickly and maintaining service availability.)
- Netflix: Running large-scale microservices and batch workloads in the cloud — Built and operated many services on AWS EC2, using automation for provisioning, scaling, and resilience across multiple availability zones. (Enabled elastic scaling and high availability for a global streaming platform.)
- Oracle: Hosting enterprise workloads and Oracle software in its own cloud — Provides OCI Compute Instances (VMs and bare metal) as the foundational compute layer for running Oracle databases, middleware, and customer applications with VCN networking and block/object storage. (Supports enterprise-grade workloads with configurable compute shapes and integration with OCI networking and storage services.)
Provider Equivalents
- AWS: Amazon EC2
- Azure: Azure Virtual Machines
- GCP: Compute Engine
- OCI: OCI Compute
Frequently Asked Questions
- What's the difference between a Compute Instance and a container (like Kubernetes)?
- A compute instance is a virtual machine: you manage the OS and can run anything on it. Containers package an app and its dependencies and share the host OS kernel, so they start faster and are lighter. Many teams run containers on top of compute instances (for example, Kubernetes worker nodes are often VMs).
- When should I use a Compute Instance?
- Use a compute instance when you need full control of the operating system, want to run software that expects a VM (legacy apps, custom agents, certain database setups), need predictable performance, or want to lift-and-shift an on-prem server to the cloud. If you only need to run small event-driven code, serverless may be simpler; if you want standardized app packaging and scaling, containers may fit better.
- How much does a Compute Instance cost?
- Pricing is typically based on the instance shape (vCPUs/OCPUs, memory, and sometimes GPU), how long it runs (per-second or per-hour billing depending on provider/service), and attached resources like block storage, boot volumes, snapshots, and network egress. Costs also vary by region and pricing model (on-demand/pay-as-you-go vs reserved/committed use). Turning instances off when not needed and right-sizing CPU/RAM are common ways to reduce spend.
Category: cloud
Difficulty: basic
Related Terms
See Also