Amazon EC2

What is Amazon EC2?

Amazon EC2 (Elastic Compute Cloud) is AWS’s core computing service. It provides virtual servers that can be launched in a matter of minutes and shut down just as quickly. Instead of waiting months for hardware to be delivered and investing in your own data centre, you rent exactly the computing power you need at any given moment – billed by the second or by the hour.

For medium-sized businesses, EC2 is often the first point of contact with the cloud. In a lift-and-shift migration, existing applications are initially moved to EC2 instances without any changes. This makes the infrastructure immediately more elastic, without the application itself needing to be adapted.

Instance types and sizing

EC2 offers hundreds of instance types, each optimised for different workload profiles. Choosing the right size – known as rightsizing – is a key factor in determining costs and performance:

• General Purpose (e.g. M-family): A balanced ratio of CPU, memory and network – ideal for web servers and application backends.
• Compute Optimized (C-family): High CPU performance for compute-intensive workloads such as batch processing or build servers.
• Memory Optimized (R family): Large amount of RAM for in-memory databases and caching.
• Storage Optimized: High local throughput for data-intensive applications.

A common mistake is oversizing: servers are chosen too large ‘just in case’. With monitoring and regular reviews, the size can be adjusted to the actual load.

Pricing models and cost optimisation

EC2 offers several billing models; combining them skilfully can significantly reduce infrastructure costs:

• On-Demand: Full flexibility with no contract commitment – ideal for unpredictable workloads and testing.
• Savings Plans and Reserved Instances: Up to 72% discount in return for a usage commitment of one to three years. Suitable for predictable baseline loads.
• Spot Instances: Up to 90% cheaper by utilising spare AWS capacity – ideal for interruptible workloads such as build jobs.

A well-thought-out FinOps strategy combines Savings Plans for the base load with Spot Instances for peak loads.

Auto Scaling and Availability

EC2 instances are rarely run individually. Using Auto Scaling Groups, AWS automatically adjusts the number of servers to match the current load. Combined with a load balancer and distribution across multiple Availability Zones, this creates a fail-safe architecture that absorbs peak loads and withstands individual hardware failures without downtime.

EC2 for SMEs: When is it the right choice?

EC2 is the right choice if you need full control over the operating system – for example, for legacy applications, specific compliance requirements or software with special licensing conditions. For new, stateless applications, container services or serverless approaches such as AWS Lambda are often more cost-effective because they reduce the administrative burden. A thorough architectural consultation clarifies in advance which service is best suited to which workload.