Appendix A — Glossary of terms

This glossary defines the acronyms and technical terms used throughout the labs. Each entry describes the term as the book uses it.

A

Access key

A pair of credentials: an access key ID and a secret access key. Together they let the AWS CLI or an SDK act as an IAM user, so treat the secret like a password.

AL2

Amazon Linux 2, the Linux distribution on the EC2 instances in this book. It uses yum for packages, useradd for accounts, and the wheel group for sudo.

AMI

Amazon Machine Image. The operating system disk image an EC2 instance boots from (similar to choosing a Docker base image).

API

Application Programming Interface. A defined way for one program to ask another for data or actions. In this book, an API usually means a web service that serves model predictions over HTTP.

ARN

Amazon Resource Name. A unique identifier for any AWS resource (for example, arn:aws:iam::123456789012:user/dev-deploy), used in policies and CLI commands.

AWS

Amazon Web Services, the cloud provider used for the server and storage labs.

B

Blob storage

Storage for files as unstructured objects (blobs) rather than as a filesystem or database. S3 is AWS’s blob storage service.

Board

A pins location where pinned objects are stored and versioned, such as a local folder (board_folder()) or an S3 bucket (board_s3()).

Bucket

The top level container for objects in S3. Bucket names are globally unique.

C

CD

Continuous Delivery (or Deployment). Automatically releasing code once it passes the CI checks.

CI

Continuous Integration. Automatically building and testing code every time a change is pushed, usually with GitHub Actions in this book.

CLI

Command Line Interface. A program you control by typing commands in a terminal, such as the AWS CLI (aws).

Container

A running instance of a Docker image. It packages an application with its dependencies and runs isolated from the host system.

CPU

Central Processing Unit, the processor. On EC2, the number of virtual CPUs is set by the instance type.

CRAN

The Comprehensive R Archive Network, the main repository for R packages.

CSS

Cascading Style Sheets, the language that controls how HTML pages look.

D

Daemon

A program that runs in the background as a service, independent of any login session. On Linux, daemons are usually managed by systemd.

DNS

Domain Name System. Translates human readable names (like example.com) into IP addresses.

Dockerfile

A text file of instructions (FROM, COPY, RUN, CMD, etc.) that Docker uses to build an image.

E

EC2

Elastic Compute Cloud. AWS’s service for renting virtual servers (instances).

EDA

Exploratory Data Analysis. Summarizing and plotting data to understand it before modeling.

EFS

Elastic File System. AWS’s managed network filesystem.

Elastic IP

A static public IPv4 address in AWS that stays the same when an EC2 instance is stopped and restarted.

Endpoint

A specific URL path and HTTP method an API responds to, such as POST /predict or GET /ping.

Environment variable

A named value set in the shell or process environment (for example, PATH or AWS_PROFILE). Used to pass configuration and credentials without hard coding them.

G

GET

The HTTP method for requesting data without changing anything on the server.

GitHub Actions

GitHub’s CI/CD service. Workflows are defined in YAML files under .github/workflows/.

GPG

GNU Privacy Guard. A tool for encrypting and signing data; package repositories use GPG keys to verify downloads.

H

HIPAA

The U.S. Health Insurance Portability and Accountability Act, which restricts how protected health information can be stored and logged.

HTML

HyperText Markup Language, the structure of a web page.

HTTP

HyperText Transfer Protocol. The request and response protocol web browsers and APIs use to communicate.

HTTPS

HTTP over TLS. The same protocol with the connection encrypted.

I

IAM

Identity and Access Management. The AWS service for users, groups, roles, and the permissions (policies) attached to them.

Image

A read only template (built from a Dockerfile) that Docker uses to start containers.

IP address

Internet Protocol address. The numeric address of a machine on a network. An EC2 instance has a private IP inside its VPC and may have a public IP.

J

JDBC

Java Database Connectivity. A standard interface Java based tools use to connect to databases.

journald

The systemd logging service. Read its logs with journalctl.

JSON

JavaScript Object Notation. A plain text format for structured data, and the format API requests and responses use in this book.

K

Key pair

A matched public key and private key. The public key can be shared freely; the private key never leaves your machine. SSH and EC2 logins rely on key pairs.

L

LLM

Large Language Model, such as Anthropic’s Claude.

Log level

The severity of a log message (DEBUG, INFO, WARNING, ERROR, CRITICAL). Setting a level filters out messages below it.

M

ML

Machine Learning.

N

NAT

Network Address Translation. Rewriting IP addresses as traffic crosses a network boundary, such as AWS’s Internet Gateway mapping a public IP to an instance’s private IP.

nginx

A web server often used as a reverse proxy in front of an application.

O

ODBC

Open Database Connectivity. A standard interface for connecting applications to databases through drivers.

OS

Operating System.

P

PAT

Personal Access Token. A token used in place of a password to authenticate with a service such as GitHub.

PATH

The environment variable listing the directories the shell searches for commands. which shows which copy of a command wins.

pins

An R and Python package for publishing and versioning objects (data, models) to a board.

POST

The HTTP method for sending data to a server, such as the inputs for a prediction.

Prototype data

An example row of the input data saved with a vetiver model, used to check that API requests have the expected columns and types.

R

RAM

Random Access Memory. On EC2, the amount of memory is set by the instance type.

renv

An R package that records a project’s package versions in a lockfile (renv.lock) so the environment can be restored.

Reverse proxy

A server (like nginx) that receives client requests and forwards them to an application behind it, often handling TLS.

RSA

A public key algorithm commonly used for SSH keys.

S

S3

Simple Storage Service. AWS’s blob storage service.

Security group

A virtual firewall for an EC2 instance. It controls which ports and IP addresses can reach the instance (ingress) and where the instance can send traffic (egress).

SSH

Secure Shell. An encrypted protocol for logging in to and running commands on a remote server.

SSL

Secure Sockets Layer, the predecessor to TLS. The name is still used loosely for TLS certificates.

Status code

The three digit number in an HTTP response that reports the outcome (for example, 200 success, 404 not found, 500 server error).

sudo

A command that runs another command with administrator (root) privileges.

Symlink

A symbolic link. A file that points to another file or directory.

systemd

The Linux service manager that starts, stops, and restarts daemons using unit files.

T

TCP

Transmission Control Protocol. The connection based transport that HTTP and SSH run over.

TLS

Transport Layer Security. Encrypts traffic between a client and a server (the S in HTTPS).

U

UI

User Interface.

URL

Uniform Resource Locator. A web address.

V

venv

Python’s built in tool for creating isolated virtual environments.

vetiver

An R and Python framework for versioning, deploying, and monitoring models as APIs.

Virtual environment

An isolated set of installed packages for a single project (venv in Python, renv in R).

VPC

Virtual Private Cloud. An isolated network in AWS where EC2 instances run.

W

WebSocket

A protocol that keeps a two way connection open between browser and server. Shiny uses WebSockets, so a reverse proxy must pass the Upgrade and Connection headers.

Y

YAML

A plain text configuration format (YAML Ain’t Markup Language) used by Quarto, GitHub Actions, and many other tools.