"Server Farm to Table"
summary
Type a URL and the browser checks caches then runs gethostbyname and /etc/hosts before UDP DNS queries that recursively ask root→.com→google.com; use dig to inspect. After DNS it opens a TCP socket (three‑way SYN/SYN‑ACK/ACK), may run TLS handshakes, then issues HTTP/1.1 requests; use curl and tcpdump to observe. Servers gzip HTML (often ~70% smaller) and send assets; browsers parse HTML into DOM, block on CSS and on scripts unless async/defer is used, honor limited parallel connections (≈6–13), then build render tree, layout, paint; HTTP/2 removes many workarounds.
AI-written summary · about
chapters (16)
- 0:00Welcome and introduction
- 0:49How URLs become IP addresses
- 3:15DNS lookup and UDP explained
- 5:25TCP handshake and connection
- 7:55TLS handshake for secure connections
- 9:02HTTP requests and server responses
- 11:56Optimizing response size with gzip
- 12:39Browser parsing and DOM construction
- 15:29Overcoming parallel download limits
- 16:55JavaScript blocking and async loading
- 17:15CSS parsing and render blocking
- 19:03Layout and painting process
- 21:53Page load events and interactivity
- 22:39Introduction to HTTP/2
- 23:01Resources and closing remarks
- 23:45Q&A: Learning resources and insights
