Server performance 7 min read Updated August 2026

Server response performance: DNS, connection, TLS, TTFB, and payload

A single total duration hides where time was spent. The report separates DNS lookup, TCP connection, TLS handshake, time to first byte, total transfer time, and response size so the slow layer can be investigated first.

At a glance

Optimize the dominant part of the request path

High TTFB often points to origin processing, cache misses, or distance; slow connection setup suggests network or TLS overhead; a large gap after the first byte points to payload size or throughput. Treat one synthetic request as a diagnostic sample, not field performance for every user.

  • Compare timing components instead of optimizing only the total.
  • Separate cold-cache, warm-cache, origin, and edge measurements.
  • Use real-user data for user experience and synthetic data for diagnosis.
  • Retest repeatedly because network and server load vary.
01

What each measurement tells you

  • DNS lookup Time to resolve the hostname; repeated delays may involve DNS provider, delegation, or resolver behavior.
  • TCP and TLS Connection and encryption setup; distance, protocol reuse, and edge placement affect it.
  • TTFB Time until the first response byte, including connection work, server processing, and network latency.
  • Download and size Time and bytes after the first byte; compression, HTML volume, and throughput matter here.
02

Find the owning layer

Repeat the request from the same region and compare a CDN cache hit with a bypassed or purged response. If hits are fast and misses are slow, profile the origin and cache policy. If both are slow, examine edge routing, TLS setup, and geography.

At the application layer, inspect database queries, external APIs, template rendering, session startup, and synchronous work. Static HTML and versioned assets can often be cached aggressively; personalized HTML requires a safer strategy.

03

Connect lab timing with user experience

Server timing influences when rendering can begin, but it is not the whole user experience. JavaScript execution, images, fonts, layout shifts, and interaction responsiveness are better assessed with browser metrics and field data.

Use the report to form a hypothesis, then confirm it with server traces, CDN analytics, Chrome tooling, and real-user measurements at relevant percentiles and regions.

Action checklist

Diagnose performance with repeatable comparisons

  1. 01Capture several measurements instead of relying on one request.
  2. 02Compare cache hits, cache misses, and a direct origin request.
  3. 03Profile server work when TTFB dominates.
  4. 04Reduce and compress HTML when transfer dominates.
  5. 05Validate the change with synthetic and real-user data.

Official references

The technical recommendations in this guide are aligned with these primary sources.