Key Terminology

• Data: We store digital data as bytes on physical media like hard drives, but in the cloud, we use systems to simplify this for us.
• Filestore: A system to save, archive, and recall specific documents and media. Usually, a file store offers low cost per unit space optimized for larger files.
• Database: A system to save and organize complex data models. They often use more advanced data structures to index and organize data for faster lookup time. Databases are compelling solutions for data models that have complicated relationships.
• Compute: Any computer operation requires some computation on a CPU (central processing unit).
• Server: A specialized computer for cloud and web services.
• Cloud Server: a computer connected to a network that is managed by someone else.
• Instance: A computational unit which may be a physical server or a virtual server that is abstracted by a cloud service.
• Cluster: A collection of instances which perform the same function. It may be used to distribute workloads.
• Autoscaling Group: A type of cluster that can increase or decrease the number of instances based on demand.
• Technical Debt: The cost (in time and resources) of additional code rework caused by choosing an easy solution now instead of using a better approach that would take longer to implement.
• Microservices: individual specialized hardware designed to accomplish a specific task. Specific tasks may include things like authentication, image processing, or data management.
• Consumable APIs: APIs that are accessible to those outside of the organization, for example, a consumable API allows programmers to search Tweets or Yelp reviews.
• Unit Tests: Ensure atomic functions and methods perform their tasks correctly or fail appropriately.
• Integration Tests: these ensure our entire endpoint perform their tasks correctly, fails appropriately, and communicates with other systems in a predictable manner (so they integrate properly).
• RAM (Random Access Memory): Data can be accessed quickly, but is erased once the server restarts. It may be okay to use RAM when prototyping, and later replace it with a database.
• Hard Drive Disk: Data remains after server restarts, but is specific to that server (not shared across servers).
• Race Condition: When an application’s behavior is dependent on other uncontrollable events. This is an issue with storing data on disks or RAM of multiple servers.
• Relational Database: can store at scale, improve search runtime, and maintain relationships between data fields.
• B-Tree: a generalization of a binary search tree, which stores sorted data, but can have more than 2 child nodes.
• Bloom Filters: a data structure that is useful for determining if an item is probably in a data set, or definitely not in the data set. Bloom filters don’t actually store the data themselves.
• Primary key: consists of one or more column in a table that are unique to each record (each row)
• Foreign key: a key in a table that contains the primary key of another table.
• Content Delivery Network (CDN): are a network of proxy servers that are placed closer to end users to deliver data and compute. CDNs reduce latency for end users.
• AES 256: Advanced Encryption Standard with a 256-bit key. This is a popular encryption standard.
• CORS: Cross Origin Resource Sharing. This defines how a client can interact with a resource, and what the client can and cannot do with that resource. Setting the CORS policy of our S3 bucket allows our client to communicate with the S3 bucket using the SignedURL pattern.
• HIPPA: HIPAA (Health Insurance Portability and Accountability Act) is a law in the U.S that requires data privacy and security for medical information.
• IAM user role: an IAM role can give a user a set of permissions to access one or more services.
• IAM service role: an IAM role gives a service a set of permissions to access one or more services.