03 / Watch shelf
YouTube
Videos and Shorts on engineering ideas, useful tools, systems, and the lessons that become clearer when shown instead of summarized.
ShortRetries are meant to improve reliability, but during an outage they can multiply traffic and prevent an API from recovering. This visual short explains retry storms, temporary failures, capped attempts, exponential backoff, jitter, and retry budgets.
ShortWebSockets are powerful, but they are not always the right choice. This visual short compares polling and WebSockets: how polling checks for updates, how WebSockets push data over a long-lived connection, and when each approach makes sense.
ShortTwo identical payment requests can reach an API at almost the same time. Without protection, both may create an order or charge the user. This visual short shows how idempotency keys, automatic retries, and an atomic database claim ensure one operation produces one saved result.
ShortRedis is only the tool. The real caching strategy is about who controls reads, writes, misses, stale data, and refreshes. This visual short explains four common caching strategies:
ShortCaching is one of the simplest ways to make APIs faster, but it happens at multiple layers. This visual explainer covers browser cache, CDN cache, server-side cache, database/query cache, cache hits, cache misses, TTL, freshness, and stale data tradeoffs.
ShortA CDN, or content delivery network, helps websites feel faster by moving static content closer to users. This visual explainer shows origin servers, edge servers, cache misses, cache hits, lower latency, reduced origin load, and faster global delivery.
ShortHow do real APIs prove identity and decide access? This visual explainer covers Basic Auth, session cookies, API keys, JWT / Bearer tokens, OAuth 2.0, OpenID Connect, RBAC, ABAC, and scopes.
ShortAuthentication and authorization are often mixed up, but they solve different problems. Authentication verifies identity: who are you?
ShortGood API design starts with HTTP. This quick explainer breaks down the core pieces behind every API request: methods, URLs, headers, bodies, status codes, cookies, CORS, caching, and protocol versions.
ShortReverse proxy and load balancer are often confused because both sit between users and servers. This quick system design explainer shows the simple difference:
ShortA load balancer sits between users and servers so one machine does not get overloaded. It receives requests first, distributes traffic across healthy servers, and reroutes traffic when a server goes down.
ShortConsistent hashing is the trick distributed systems use to add or remove servers without remapping almost every key. In this short, we compare normal modulo hashing with hash rings, clockwise lookup, server failures, and virtual nodes for better load balancing.