IBM Full Stack Software Developer flashcards
IBM Full Stack Software Developer · Build recall with a focused review session.
33 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Git commit object
- A commit stores a pointer to a tree (the snapshot), pointers to parent commits, author and committer data, and a message. Its SHA-1 hash identifies it uniquely.
- Idempotent HTTP methods
- GET, PUT, DELETE, and HEAD produce the same server state when repeated. POST and PATCH are not idempotent, so retried requests may duplicate side effects unless idempotency keys are used.
- REST statelessness
- Each request carries all information the server needs to process it. The server keeps no client session state between calls, which simplifies scaling and lets any instance handle any request.
- Agile user story
- A short requirement written as role, goal, and benefit: as a user I want a capability so that I get value. Acceptance criteria define when the story is complete.
- CSS specificity order
- Inline styles beat IDs, IDs beat classes, attributes and pseudo-classes, which beat element and pseudo-element selectors. Later rules win only when specificity ties.
- Flexbox versus Grid
- Flexbox arranges items along a single main axis with wrapping. Grid defines explicit rows and columns, positioning items in two dimensions at once.
- JavaScript event loop
- The call stack runs synchronous code; finished microtasks (promise callbacks) drain before the next macrotask (timers, I/O callbacks). This ordering explains async output sequences.
- box-sizing border-box
- With border-box, a declared width includes content, padding, and border, so adding padding does not enlarge the element. The default content-box excludes padding and border from width.
- React one-way data flow
- Props pass data from parent to child and are read-only in the child. A child requests changes by invoking callback props, keeping state ownership predictable.
- Fetch response.ok
- The fetch promise rejects only on network failure. HTTP 404 or 500 responses still resolve, so code must check response.ok or the status before parsing the body.
- Express middleware chain
- Middleware functions receive request, response, and next. Calling next passes control onward; calling next with an error skips to error-handling middleware with four parameters.
- ACID transactions
- Atomicity means all statements commit or none do. Consistency enforces constraints, isolation prevents interfering concurrent reads, and durability preserves committed data across crashes.
- SQL INNER versus LEFT JOIN
- INNER JOIN returns only rows matching in both tables. LEFT JOIN returns every left-table row, filling right-side columns with nulls where no match exists.
- JWT payload confidentiality
- A JSON Web Token's payload is base64url-encoded, not encrypted. Anyone holding the token can decode its claims, so secrets and sensitive personal data must stay out.
- Password hashing with salt
- Store passwords as slow salted hashes such as bcrypt or Argon2. The salt defeats rainbow tables, and the slow work factor makes brute-force guessing expensive.
- Database index tradeoff
- Indexes accelerate lookups and sorts but add storage and slow inserts, updates, and deletes because each index must be maintained. Index columns used in filters and joins.
- Node.js non-blocking I/O
- Node runs JavaScript on one thread and delegates I/O to the system, resuming via callbacks. CPU-bound work blocks every request, so it belongs in worker threads or separate services.
- IaaS PaaS SaaS
- IaaS supplies virtual machines and networking, PaaS supplies managed runtimes and middleware, SaaS supplies complete applications. Customer control shrinks from IaaS to SaaS.
- Availability zone resilience
- A region contains isolated availability zones with independent power and networking. Deploying replicas across zones keeps a service running when one data center fails.
- Horizontal versus vertical scaling
- Horizontal scaling adds more instances behind a load balancer; vertical scaling gives one machine more CPU or memory. Stateless services scale horizontally without code changes.
- Cloud shared responsibility
- The provider secures physical facilities, hardware, and managed-service internals. The customer secures data, identities, access policies, and configuration of what they deploy.
- Serverless cold start
- A serverless function that has scaled to zero needs initialization on the next invocation, adding latency. Provisioned concurrency or keeping instances warm reduces this delay.
- Container versus virtual machine
- Containers share the host kernel and isolate processes, so they start fast and stay small. Virtual machines run a full guest operating system with stronger isolation and more overhead.
- Docker image layers
- Each Dockerfile instruction adds a read-only layer. Builders reuse cached layers while instructions and inputs are unchanged, so ordering instructions well speeds rebuilds.
- Container ephemeral filesystem
- Writes inside a running container live in a writable layer discarded on removal. Persistent data must go to a named volume or an external service.
- Docker port publishing
- The -p host:container flag maps a host port to a container port so external clients can connect. Without publishing, the container port is reachable only inside its network.
- Kubernetes Pod and Deployment
- A Pod is the smallest deployable unit, holding one or more containers. A Deployment manages ReplicaSets to keep the desired pod count and perform rolling updates.
- Docker Compose purpose
- A YAML file declares multiple services, networks, and volumes for one application. One command starts the whole stack locally, mirroring production service wiring.
- Continuous integration pipeline
- Every commit triggers automated build, lint, and tests. Fast feedback catches integration defects early, and a green build produces a versioned artifact for deployment.
- Blue-green deployment
- Two identical environments exist; the idle one receives the new release, is verified, then traffic switches over. Rollback means switching traffic back to the previous environment.
- Canary release
- A new version reaches a small slice of traffic first. Metrics are compared against the stable version, and the rollout expands only if health stays acceptable.
- SLI SLO and SLA
- An SLI is a measured indicator such as latency or error rate. An SLO is an internal target for that indicator; an SLA is a contract with customer-facing consequences.
- Infrastructure as Code idempotence
- Declarative templates describe desired resources. Applying the same template repeatedly converges to the same state without duplicating resources, enabling safe automated provisioning.