An Overview of ~50 AWS services

1. TL;DR: What happened?

Found this really cool video offering a quick presentation of 50 AWS services (Out of 200+!).

I thought it would be nice to have a cheatsheet to help remember the content. So here’s a mind map. You can also read the text content below, if it interests you.

2. Resources

• Fireship video about AWS services
• Be A Better Dev video about AWS
• Interesting read about Kinesis vs Kafka

1. Hosting an application

1. Elastic Cloud Compute (EC2): Create virtual instance on cloud, choose OS, memory, computing power and rent. Common to use for web application.
2. Elastic Load Balancing: Distribute traffic to multiple EC2 instances.
3. Cloud Watch: Collect logs and statistics about EC2 instances.
4. Auto Scaling: Automatically create new instances as needed.
5. Elastic Beanstalk: Further abstraction of EC2, Load Balancing, Cloud Watch and Autoscaling. Choose a platform and let everything happen automatically.
6. Lightsail: Easier and lazier version of Beanstalk.
7. Lambda: Upload a block of code to be executed, and call/pay when you need.
8. Serverless Repo: Basically github for lambda.
9. Elastic Docker Registry: Upload your Docker image.
10. Elastic Container Service: Pull down image from Elastic Docker Registry and run image on EC2 instances.
11. Fargate: Automatically manage allocations of EC2 instances for Docker images.

2. Simple Data storage

1. Simple Storage Service: Basically Google Drive.
2. Glacier: Store more for cheaper, but at the cost of retrieval speed.
3. Block Store: Much faster, but you have to configure yourself.
4. Elastic File System: Everything managed, but get ready to pay $$$.

3. Databases

1. Non-relational databases

   1. SimpleDB: Oldest and simplest
   2. DynamoDB: Document-based, basically succcessor to SimpleDB.
   3. DocumentDB: Copy of MongoDB to avoid licensing issues. 1:1 mapping.

2. Relational databases

   1. Relational Database Service (RDS): Service supporting all major SQL-based databases, such as PostgreSQL, MySQL, MariaDB, Oracle, Microsoft SQL Server.
   2. Aurora: Amazon proprietary version of MySQL, claims to have throughput 5X of MySQL and 3X of PostgreSQL. Migration from Oracle and Microsoft SQL Server seems harder.
   3. Aurora Serverless: Pay-as-you-go, automatic scaling version of Aurora.

3. Other

   1. Neptune: Graph database, great for highly related data like recommendations, social network.
   2. Elastic Cache: Fully managed version of redis, which is an in-memory database offering much faster speed.
   3. Timestream: Database made for time series data like stocks with added functinality.
   4. Quantum Ledger: Ledged database providing a transparent, immutable and verifiable transaction log.

4. Analytics

1. Data Extracting, Transforming, and Loading
   1. Redshift: Popular structured data dump for enterprises pending later analysis. (Apparently it means shift away from Oracle, which is red. Legends…)
   2. Lake Formation: Data lake for unstructured data.
   3. Kinesis: Use to capture real-time data streams.
   4. Amazon Managed Streaming for Apache Kafka (MSK): Alternative to Kinesis which allows using native Kafka APIs to populate data lakes, stream changes etc.
   5. Glue: Manager to handle/automate ETL without programming
2. Machine Learning Models
   1. Sagemaker: IDE for meachine learning, including prepare, build, train/tune and deploy (This is probably worth investigating more)/
   2. Rekognition: Image and video analysis using pre-trained model. Custom models starting with n=10 labeled-images.
   3. Lex: Building chatbots, Natural Language Understanding focused API/

5. Additing functionalities to applications

1. Identity and Access Management (IAM): Manage access to AWS services in your organization.
2. Cognito: Add sign-up, sign-in to web/mobile apps quickly.
3. Simple Notification Service: Send push notifications to your users.
4. Simple Email Service: Send email notifications to your users .