AWS Compute Services: EC2 Instance Types

Choosing the right Amazon EC2 instance type is a critical skill for the SAA-C03 exam. Think of EC2 instances like choosing a vehicle for a specific job: you wouldn’t use a Vespa to move a house, and you wouldn’t use a semi-truck to deliver a single envelope. AWS categorizes instances based on their hardware capabilities—CPU, Memory, Storage, and Networking—to ensure you get the best price-performance ratio for your specific workload.

The Real-World Analogy

• General Purpose (M/T): Your reliable family sedan. Good for commuting, grocery shopping, and most daily tasks.
• Compute Optimized (C): A high-performance sports car. Built for speed and processing power, but lacks trunk space (memory).
• Memory Optimized (R/X): A massive moving van. It can carry a huge amount of data in its cargo hold (RAM) for processing.
• Storage Optimized (I/D/H): A heavy-duty freight truck. Designed for high-speed local data access and massive storage throughput.

Core Instance Families

1. General Purpose (Balance)

Used for diverse workloads where CPU, memory, and networking are balanced.

• T-Series (Burstable): Uses “CPU Credits.” Ideal for low-traffic websites or small dev environments. If you run out of credits, performance is throttled to a baseline.
• M-Series: The “M” stands for “Main.” Used for web servers, small databases, and enterprise applications.

2. Compute Optimized (C-Series)

Ideal for compute-bound applications that benefit from high-performance processors.

• Use Cases: Batch processing, media transcoding, high-performance web servers, scientific modeling, and dedicated gaming servers.

3. Memory Optimized (R, X, High Memory)

Designed to deliver fast performance for workloads that process large data sets in memory.

• R-Series: The “R” stands for “RAM.” Perfect for high-performance databases (MySQL, PostgreSQL) and distributed web-scale in-memory caches (Redis).
• X-Series: For massive SAP HANA workloads or big data processing.

4. Storage Optimized (I, D, H)

Designed for workloads that require high, sequential read and write access to very large datasets on local storage.

• I-Series (IOPS): Uses NVMe SSDs. Best for NoSQL databases (Cassandra, MongoDB).
• D-Series (Density): Best for Massively Parallel Processing (MPP) and MapReduce.

Comparison Table: Choosing the Right Family

Decision Matrix: If-Then Guide

• If you need to run a high-performance database like SQL Server, choose R-Series.
• If you are performing video encoding or scientific modeling, choose C-Series.
• If you need local NVMe storage for a NoSQL cluster, choose I-Series.
• If you have a small app with occasional spikes in traffic, choose T-Series.
• If you need to run Machine Learning or Graphics rendering, choose G or P (Accelerated Computing).

Exam Tips and Gotchas

• The Nitro System: Modern instance types (C5, M5, R5, etc.) use the Nitro Hypervisor, which provides better performance by offloading networking and storage functions to dedicated hardware.
• Burstable Performance: On the exam, if a scenario mentions “unexpectedly poor performance” on a T2/T3 instance, the answer is likely that the instance has exhausted its CPU Credits.
• Instance Store vs. EBS: Remember that storage-optimized instances often use “Instance Store” (ephemeral) storage. If the instance stops, the data is lost. Always pair with EBS for persistence unless the application handles replication.
• ARM vs. Intel: Graviton processors (prefixed with ‘g’, e.g., M6g) provide the best price-performance but require your application to support the ARM architecture.

Topics covered:

Summary of key subtopics covered in this guide:

• Instance Family Mnemonics (M, C, R, T, I, D, G, P).
• Burstable performance and CPU Credit mechanisms.
• Workload-to-Instance mapping for optimization.
• Hardware features: Nitro System, NVMe, and Graviton.
• Selection criteria based on CPU, Memory, and Storage requirements.