Privacy and anti-tracking solutions

Hardware Virtualization + Restore Mode — Virtualize the ID, Not the Machine: No Performance Loss

How do software and websites identify a computer? From hardware fingerprints like disk serial numbers, NIC MAC addresses, and motherboard/BIOS machine codes, to system identifiers like Windows machine codes, volume serial numbers, and OS installation IDs, to browser Canvas/font/timezone fingerprints, public IP addresses, and the accumulated data of cookies, cache, login states, and logs—dozens of identifiers are cross-referenced. Changing accounts, clearing cache, or altering individual parameters won't evade it. vDisk uses hardware virtualization to virtualize this entire 'identification layer' as a whole. Software reads only virtual IDs, while the OS still runs directly on the local CPU/GPU, not in a VM. Combined with a restore mode, all traces are deleted and reset to zero upon reboot. Two layers combined: real fingerprints cannot be collected, historical traces cannot be left, and performance remains equivalent to a physical machine.

Today's tracking is far more than just one serial number

A single computer has dozens of readable identifiers; trackers cross-reference them, and even if you change accounts, clear the cache, or alter one or two of them, the remaining ones can still match you.

Hardware fingerprint

Disk serial number, NIC MAC address, motherboard SMBIOS / UUID, BIOS information, CPU model and serial number. These are burned into the hardware, stable, unique, and nearly impossible to alter at the software layer, making them the hardest form of 'device ID'.

System and firmware identifiers

Windows machine codes: MachineGUID, machine SID, Volume Serial, system installation ID, product key. These persist even after software reinstallation or account changes; many licenses and risk controls are tied to them.

Network identifier

Public IP, NIC MAC, LAN hostname, gateway, and DNS characteristics. The network fingerprint of the same egress point and the same machine is often used to determine 'whether it is the same person / same batch of devices.'

Browser and software fingerprinting

Canvas/WebGL rendering fingerprint, installed font list, User-Agent, screen resolution, timezone/language, and license machine codes read by desktop software. Even without cookies, these can be combined to form a highly unique "fingerprint."

Storage residual traces

Cookies, LocalStorage, browsing and download history, login tokens, cache and temporary files, registry leftovers, logs. On public terminals, one person's traces are directly visible to the next.

Cross-correlation identification

The real challenge isn't individual identifiers, but 'association': as long as any of the stable identifiers above remains unchanged, multi-instance detection, cross-account linking, and device fingerprinting can re-identify you—so making piecemeal changes to a few items is useless.

Common 'anti-tracking' methods, why are none of them thorough?

The problem is that each only blocks one layer, while detection is based on cross-identification of multiple markers — missing one layer renders all previous efforts futile.

Piecemeal approach, always some slip through

  • Manually changing MAC/serial number: can only modify a few items at the software layer; motherboard UUID, volume serial number, BIOS, etc. cannot be changed, often reset on reboot, and cross-validation by software quickly exposes discrepancies.
  • Clearing cookies / opening incognito: only removes traces at the browser level; hardware fingerprints and system machine codes remain untouched, desktop software still recognizes them.
  • Fingerprint browsers: only disguise within the browser; local software still reads real hardware. Moreover, if disguise parameters conflict, it becomes even more conspicuous.
  • Installing a VM: full machine simulation leads to performance degradation, graphical lag, and often detected as a virtual environment by software, causing it to refuse to run.
  • Manual processing per device: managing dozens or hundreds of terminals is simply unmanageable; once configurations become inconsistent, it actually exposes 'this batch of machines has issues'.

vDisk's Two-Layer Defense

  • Hardware virtualization virtualizes the entire 'identification layer' at the system bottom: disk serial number, MAC, motherboard UUID, machine code, etc. all become virtual IDs; browsers and local software read these.
  • Virtual IDs can be configured uniformly per terminal or independently per device, issued as sets without conflict; real hardware fingerprints cannot be captured from the source.
  • Restoration mode: All changes are written to an incremental layer, deleted upon reboot—returning to a clean system in 30 seconds. Traces such as cookies, cache, login states, logs, etc., are all reset to zero.
  • The system still runs directly on the local CPU/GPU — not a virtual machine, performance equals a physical machine, large software and graphics applications run normally.
  • One master image for the entire computer lab, one backend to distribute uniformly, anti-tracking policies take effect in batches, consistent across the entire venue.

Virtual ID and virtual machine are not the same thing at all

Virtual machines simulate an entire computer within software; vDisk only virtualizes the identification layer, the system still runs on real hardware

Comparison Dimensions Traditional virtual machine vDisk Virtual ID
Where the system runsInside a Hypervisor virtual machine, running through a layer of software emulation.Terminal's local CPU/GPU runs directly, performance equal to a physical machine.
Performance lossCPU / Memory / GPU performance is generally discounted, with noticeable lag in graphics applications.CPU 损耗 <3%,大型软件、图形课程照常跑
What is virtualizedFull hardware simulation is costlyOnly virtualizing the identification layer like disk serial number, MAC, machine code, etc.
Software compatibilityOften detected as a virtual machine environment, some software restricts execution.Real physical machine environment, software installs and runs normally
Trace cleanupRequires manual deletion of snapshots/clones, making management cumbersomeRestore mode: clean on reboot, policies uniformly deployed across the entire computer room

Which scenarios need 'cannot collect, cannot retain'?

Privacy for public terminals, cleanliness for exam environments, diversity for test environments — three uses of the same mechanism.

Shared labs and public terminals

Libraries, electronic reading rooms, training classrooms: one user's account, browsing history, and cache are completely invisible to the next user. A restart restores a fresh, clean environment, ensuring user privacy and mutual non-interference.

Exam and assessment lab

Examination environments are uniformly deployed in a closed state, ensuring consistency between candidates with no residual history; one-click restoration after the exam, so any actions from a previous session do not affect the next, guaranteeing fair and reviewable environments.

Software testing and compatibility validation

Switch virtual hardware IDs and system environments on demand for the same batch of terminals to test activation, licensing, and binding logic; restore after testing, environments are disposable.

Test in a real lab: no fingerprints left, no traces retained, no performance compromise.

Apply for a free trial: use your existing terminals to verify the virtual ID read results, restore speed, and performance of large software — let the data speak.

Last updated: 2026-06-30