Browser profile with aligned network information measurements and one documented signal change

Network Information Signal Drift: Check Connection Type, RTT, Downlink, and Data Saver Before Browser Profile Reuse

Network information signal drift occurs when a reusable browser profile reports a different connection type, effective connection class, round-trip estimate, downlink estimate, or data-saver state without a documented network, device, browser, or policy change. Capture these values under matched conditions before deciding whether a profile is ready for reuse.

These fields are estimates exposed by the browser when the Network Information API is supported. They may be rounded, unavailable, or updated as conditions change. Treat them as environment-consistency evidence, not as proof of identity, location, anonymity, or account outcome.

Record The Complete Network Information Set

FieldWhat it describesWhat to record
typeReported physical connection category when implementedValue, support status, device, and session context
effectiveTypeBroad performance class such as slow-2g, 2g, 3g, or 4gValue, test phase, and route conditions
rttRounded round-trip estimate in millisecondsRaw value and a short settled sample series
downlinkRounded effective downlink estimateRaw value, unit, and sample timing
saveDataUser or system preference to reduce data useBoolean state and owning OS/browser setting
change eventNotification that exposed network information changedTimestamp, old set, new set, and the event context

1. Confirm Support Before Comparing Values

Record whether navigator.connection, navigator.mozConnection, or navigator.webkitConnection exists. Unsupported fields and explicit null values are not the same as a changed value. Keep support status in the profile environment baseline so future captures use the same interpretation.

2. Match The Workstation, Browser, And Session

Use the same workstation, operating-system build, browser build, profile configuration, launch mode, power state, VPN state, and remote-session state. Compare the workstation and browser baseline before treating a field difference as profile drift.

3. Separate Route Changes From Profile Changes

Record the local network, proxy endpoint, exit address, protocol, and route owner. Network Information API values can respond to the client connection and browser estimates rather than the public exit alone. Keep the proxy route baseline fixed while comparing profiles.

4. Capture Initial And Settled Samples

Capture once immediately after launch, then again after the profile and target page have been idle under a fixed route. Record several bounded samples rather than one number. A pass is a stable pattern within the expected rounded range. A single early change is ambiguous until the settled capture is complete.

5. Interpret RTT And Downlink As Rounded Estimates

Do not compare rtt or downlink as laboratory measurements. Browsers may round values, use recent observations, or apply privacy protections. Compare ranges and update behavior under matched conditions. Repeated large changes with no route or workstation explanation deserve review; small rounded differences may be expected.

6. Check Effective Type Without Inferring Physical Hardware

effectiveType is a performance class, not a declaration that the device uses a particular radio. A desktop connection can still be classified by observed performance. Record the raw value and conditions without translating it into unsupported claims about the physical network.

7. Verify Data Saver At Its Owning Setting

If saveData changes, check the operating-system and browser data-reduction settings, managed policies, and session type. Do not force the value merely to match an old capture. Document the approved preference and update the baseline when the owner intentionally changes it.

8. Listen For Change Events During A Bounded Window

Attach one change listener during a short controlled observation window. Record the full old and new field sets with timestamps. Avoid continuous background polling. A change event is useful only when it can be tied to a route, interface, policy, power, or session event.

9. Relaunch One Variable At A Time

Use a controlled profile relaunch with the same workstation and route. If the settled set returns to baseline, the initial mismatch may have been load timing. If it remains different, change only one approved variable next: browser build, route, OS preference, or session type.

10. Compare Raw Evidence Before Normalizing

Store raw support status and raw values first. A normalized summary can group rounded RTT or downlink values for review, but it must not erase whether a field was missing, zero, false, or unsupported. The advanced fingerprint review should retain both layers.

Reuse Decision Table

Observed resultDecisionNext action
Support and settled values match the approved baselineReuseRecord the capture and proceed gradually
Small rounded RTT or downlink differences onlyReview rangeCompare the bounded sample series
Change follows a documented route or network eventExpected changeUpdate the baseline with owner approval
Support disappears after a browser or policy updatePauseVerify implementation and managed policy
Connection class or data-saver state changes without explanationStop reusePreserve evidence and investigate one variable at a time

11. Keep A Compact Evidence Record

Record the profile ID, workstation, OS build, browser build, launch mode, local network, proxy route, support status, type, effective type, RTT, downlink, data-saver state, change events, initial and settled timestamps, expected explanation, owner, and reuse decision. Add it to the profile reuse checklist.

12. Stop When Conditions Cannot Be Matched

Pause reuse when the workstation, browser build, route, power state, remote-session state, or managed policy cannot be confirmed. Do not manufacture a preferred value, hide a real network change, or claim that one matching field makes an environment safe. Consistency requires a coherent set and an auditable explanation.

Use Matched Conditions, Ranges, And Ownership

A reliable network information review checks support first, fixes the workstation and route, captures initial and settled values, compares rounded ranges, records change events, and traces each difference to an owner. The goal is not one universal network signature. It is a repeatable browser-profile environment whose changes are expected, documented, and reviewable.