16 articles about system-design-concepts
HEREâS WHAT I LEARNED BUILDING A SUBSCRIPTION MANAGEMENT SYSTEM FROMÂ SCRATCH. If youâve ever built an app, youâve probably started the same way I did a fronten...
WHY COMMON ASSUMPTIONS ABOUT SCALING LEAD TO FRAGILEÂ SYSTEMS Scaling is often seen as a technical problem. More users arrive, and the system needs to handle i...
UNDERSTANDING THE GAP BETWEEN ACTUAL PERFORMANCE AND PERCEIVED LATENCY In previous parts [https://medium.com/@akshatjme/observability-you-cant-fix-what-you-can...
UNDERSTANDING SYSTEM BEHAVIOR BEYOND LOGS AND DASHBOARDS In previous parts, we explored how systems fail under load and how design decisions influence performa...
WHY TESTING FOR STABILITY OFTEN HIDES THE REAL LIMITS OF YOURÂ SYSTEM In previous parts, we explored how systems behave under pressure. Load testing is meant t...
80% OF USER PROBLEMS ARE REPEATED PATTERNS. SO WHY ARE WE SOLVING THEM MANUALLY EVERYÂ TIME? If youâve ever built a help and support system, youâve probably don...
HOW TO BUILD BACKEND SYSTEMS THAT CONTINUE TO WORK EVEN WHEN THINGS GOÂ WRONG In earlier parts [https://medium.com/@akshatjme/list/the-hidden-reasons-your-backe...
WHY CONTROLLING TRAFFIC MATTERS MORE THAN HANDLINGÂ IT In Part 1 [https://akshatjme.medium.com/why-your-backend-stops-performing-overnight-2d5e3a2f263f], we saw...
WHY MOST BACKEND PERFORMANCE ISSUES EVENTUALLY LEAD BACK TO THEÂ DATABASE In Part 1 [https://akshatjme.medium.com/why-your-backend-stops-performing-overnight-2d...
MOST BACKEND SYSTEMS DO NOT FAIL DURING DEVELOPMENT. They fail later, in production, when everything seems to be working fine. One day the system is fast and...
Cloud storage systems are not accessed by trustâââthey are accessed by authorization. Modern applications cannot safely embed permanent credentials in frontend...
HOW SYSTEMS PROTECT THEMSELVES FROM TOO MANY REQUESTS Modern systems donât fail because they are badly written. They fail because they receive more requests t...
In the previous two articles, we saw something important: * To scale reads, we reduce the work the database does per query. [LINK] [https://medium.com/@akshat...
Distributed systems do not fail because of computation. They fail because of communication. When services talk to each other directly, they become tightly coupl...
In the previous article [https://medium.com/@akshatjme/5671a7ac80e1], we saw how to scale reads by reducing the amount of work the database has to do for every...
Learn how to scale database reads as traffic grows through replicas, caching, materialized views, denormalization, and other proven techniques that reduce repeated work.