Skip to content

LABS: Learning Segmentation by Observation

These labs are designed to help you see and feel how segmentation works in practice. You do not need deep networking knowledge to follow them.

The goal is observation, not performance testing or traffic bypassing.

These labs focus on observing delivery behavior. They do not change content and do not inspect encrypted traffic. For HTTPS, segmentation applies only to CONNECT tunnels.

Before You Start

Requirements

  • SegmentedProxy running locally
  • Basic familiarity with editing rules.txt
  • A browser or curl

What to Watch

During these labs, pay attention to:

  • proxy logs
  • timing differences
  • how data appears to arrive

You are not expected to inspect packets or decrypt traffic.

Lab 1: Fixed Segmentation vs No Segmentation

Goal

Understand how fixed-size chunks change delivery behavior.

Setup

  1. Start with no segmentation rule for a test domain.

  2. Make a request and note:

  3. how fast the response appears

  4. how logs look

  5. Now add a fixed segmentation rule:

*.example.com=segment_upstream,strategy=fixed,chunk=256

Observe

  • The response may appear in small steps
  • Logs may show multiple forwarded chunks
  • The content is the same, but the delivery feels different

What You Learn

Segmentation affects how data arrives, not what data arrives.

Lab 2: Adding Delay Between Chunks

Goal

See how timing affects applications.

Setup

Modify the rule:

*.example.com=segment_upstream,strategy=fixed,chunk=256,delay=50

Observe

  • Pages may load more slowly
  • Text or data may appear gradually
  • The browser may show loading indicators longer

What You Learn

Small delays can strongly influence perceived speed, even without changing bandwidth.

Lab 3: Random vs Fixed Segmentation

Goal

Compare predictable vs unpredictable chunking.

Setup

Use a random strategy:

*.example.com=segment_upstream,strategy=random,min=128,max=1024,delay=10

Observe

  • Chunk sizes vary
  • Timing feels less regular
  • Logs show uneven forwarding

What You Learn

Different segmentation strategies create different delivery patterns.

Lab 4: HTTPS and CONNECT-only Segmentation

Goal

Understand where segmentation applies for HTTPS.

Setup

Use a CONNECT-only rule:

*.example.com=segment_upstream,scheme=https,method=CONNECT,strategy=fixed,chunk=512

Observe

  • Only HTTPS traffic is affected
  • HTTP requests remain unchanged
  • Encrypted content is never inspected

What You Learn

For HTTPS, segmentation applies to the tunnel, not the content.

Lab 5: When Segmentation Seems to Do Nothing

Goal

Learn common reasons why effects are not visible.

Try This

  • Use a very large chunk size
  • Use a fast, small response
  • Remove delay

Observe

  • No visible difference
  • Fewer log entries

What You Learn

Segmentation effects depend on:

  • data size
  • timing
  • rule matching

Sometimes “nothing happens” is the expected result.

Key Takeaways

  • Segmentation changes delivery behavior, not content

  • Effects are easier to see with:

  • larger responses

  • smaller chunks
  • added delays
  • HTTPS segmentation works only through CONNECT
  • Rules control where segmentation applies

If segmentation effects feel subtle at first, that is normal. This project is designed to teach observation and understanding, not to produce dramatic visual changes.