Easy concepts in Databases & Architecture, page 2
Key-Value Store: The Simplest Database Model
A key-value store is a giant dictionary. You give it a unique key, like "user:123", and it returns the associated data. It's the foundation for caching and session management. The footgun is trying to query by value—it's built for key lookups only.
Document Databases: Store Data as Flexible Objects
A document database stores data as self-contained objects, like JSON, instead of rows and columns. It's ideal for user profiles or product catalogs where each item might have different attributes. The footgun is treating it as a schema-less free-for-all.
Read Replicas: Scale Out Your Database Reads
A read replica is a read-only copy of your database that handles query traffic. Use it for read-heavy apps to prevent your primary DB from becoming a bottleneck. The footgun: replication is asynchronous, so reads from a replica can return slightly stale data.
Differential Backups: Faster Backups, Simpler Restores
A differential backup saves all changes since the last full backup, making daily backups faster. To restore, you only need the full backup and the latest differential file. The footgun: each differential file grows larger until the next full backup is made.
Role-Based Access Control (RBAC) in Databases
RBAC bundles permissions into roles, like 'analyst' or 'admin', instead of assigning them to individuals. This simplifies managing who can read or write data in a database. The footgun is creating too many roles, making it as complex as individual permissions.
Connection Pooling: Don't Re-Open, Reuse
A connection pool is a valet service for database access. Instead of creating a new connection for every request, you borrow a ready-made one and return it. This avoids costly setup/teardown in web apps.
JDBC: Java's Universal Translator for Databases
JDBC is Java's universal adapter for databases, letting your app speak SQL to any database via a standard API. It's used for connecting, querying, and managing transactions. The biggest footgun is building SQL strings directly; always use PreparedStatements.
ODBC: The Universal Translator for Databases
ODBC acts as a universal translator, letting one application speak to many different relational databases. Your app uses the standard ODBC interface, and a specific "driver" handles the unique protocol for each database. The footgun is performance overhead.
Connection String: Your App's Key and Address to Data
A connection string is your app's address and key to a data source. It bundles the host, port, database name, and credentials into a single string for a driver to use. The main footgun is committing credentials to version control by hardcoding the string in.
Database as a Service (DBaaS): Rent, Don't Build
DBaaS is like leasing a database instead of owning it. A cloud provider handles the backups, patching, and scaling, so you can focus on your app. The main footgun is assuming "managed" means you can ignore configuration, query performance, and costs.
Amazon RDS: Managed Relational Databases in the Cloud
Amazon RDS is like hiring a DBA to manage your database's plumbing. It's for when you need a SQL database like PostgreSQL or MySQL without the hassle of patching and backups. The footgun is assuming it's 'serverless'—you still manage cost and performance.
Data Pipelines: From Raw Data to Actionable Insights
A data pipeline is the plumbing for your data, moving it from raw sources to a refined state for analysis. It feeds dashboards and ML models by cleaning data from APIs and databases. The key footgun is choosing batch processing for real-time needs.
Message Queues: Decoupling Your Services
Think of a message queue as a digital post office for your services. It lets one part of your system drop off a task for another to handle later, decoupling them so they don't have to run in lock-step.
Inverted Index: How Search Engines Find Your Keywords
An inverted index is like a book's index: it maps keywords to the documents containing them. This is the core of full-text search in search engines and databases, allowing instant lookups.
Cache-Aside Pattern: Your App Owns the Cache
The Cache-Aside pattern makes your application the gatekeeper for the cache. On a read, your code checks the cache first; on a miss, it fetches from the database and writes to the cache. This speeds up read-heavy apps. The key footgun is stale data.
Vector Embeddings: Turning Meaning into Math
Vector embeddings turn complex data like words or images into lists of numbers (vectors). This lets computers measure "similarity" by calculating the distance between these vectors, powering search and recommendations.
Data Retention Policy: Your Schedule for Deleting Data
A data retention policy is your company's official schedule for deleting data, not a plan to keep it forever. It's essential for legal compliance (like GDPR) and managing storage costs.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles