Datasets:
The dataset viewer is not available for this split.
Error code: RowsPostProcessingError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Installation Is Not Authorisation: Execution-Finality Architecture for DMA Article 6(4) Alternative App Distribution
Author: Sangam Kumar Das Affiliation: Independent Inventor, Balasore, Odisha, India Part of: DAS Protocols Patent Family — Execution-Finality / Candidate-Act Architecture Status: Prior-art / technical disclosure record (machine-indexable repository entry) License: CC BY-NC 4.0 (Creative Commons Attribution-NonCommercial 4.0 International) Contact: sangamdas@gmail.com
Abstract
Article 6(4) of the EU Digital Markets Act requires gatekeepers to permit installation and effective use of third-party applications and alternative app stores, while still allowing strictly necessary and proportionate security measures. This creates a technical problem deeper than installation itself: permission to install software should not automatically mean authority for every subsequent machine-generated act of that software to become effective.
This disclosure presents an execution-finality architecture that technically separates two distinct questions that existing security controls (code signing, sandboxing, permissions, authentication, OAuth) conflate:
- Distribution Authority — May this application be installed and run?
- Execution-Finality Authority — May this particular operation become externally effective?
Every security-sensitive operation (network export, file transmission, payment, sensor access, protected-state change) is first held as a non-final Candidate Act. Before it can produce an external consequence, a Protected Enforcement Domain validates identity, purpose, destination, scope, freshness, and revocation state. Only then is a narrowly scoped, non-bearer execution authority released and independently verified at the Finality Sink — the boundary where the act would first become real.
The architecture is explicitly deployable using present-day technology (existing OS kernel mediation points, TEEs, ARM TrustZone, Secure Enclave-class processors, TPMs, secure elements) and does not require new hardware, a new internet protocol, or blockchain consensus. It is intended to sit alongside, not replace, existing app-store review, code signing, sandboxing, and permission systems.
Core Inventive Principle
A machine may compute, generate, sign, or route an act, but computation alone does not create authority for that act to become externally effective.
Applied to DMA Article 6(4):
RIGHT TO INSTALL ≠ UNRESTRICTED RIGHT TO EFFECTUATE
Distribution source and execution authority are technically decoupled. Applications from an official store and applications from a DMA-enabled alternative store encounter the same neutral consequence-boundary enforcement, so open distribution does not require unrestricted execution authority.
Key Terms (Locked Vocabulary)
| Term | Meaning |
|---|---|
| Candidate Act | A proposed operation (network export, payment, file share, sensor access, etc.) generated by an application but not yet permitted to produce its external effect. |
| Non-Effective State | The holding state of a Candidate Act — computationally present, but incapable of producing a protected consequence. |
| Protected Enforcement Domain | The isolated validation environment (software or hardware-backed) that checks identity, purpose, destination, scope, freshness, and revocation before authorising an act. |
| Finality Sink | The functional boundary — network stack, file-export service, payment terminal, IPC broker, device controller, etc. — where a Candidate Act would first become externally effective. |
| Execution Handle / Scoped Authority | A narrowly bound, short-lived, preferably non-bearer capability tying validated authority to one specific Candidate Act, destination, and validity window. |
| LAVR (Ledger Anchored Validation Receipt) | Protected evidence generated at successful validation, associated with the enforcement event rather than functioning as mere post-hoc logging. |
What This Architecture Is Not
- Not a replacement for app-store review, malware scanning, code signing, or sandboxing — it adds one further check at the point of consequence.
- Not a claim that all future hardware is required — an initial deployment can be predominantly software-based using existing OS mediation points, strengthened later by already-available TEEs/secure elements.
- Not a claim of measured sub-millisecond latency — hot-path enforcement is described as an engineering target using bounded local verification (digest match, nonce/replay check, expiry, epoch check), not a benchmarked figure.
- Not a system that disables an entire application or store on failure — enforcement fails closed at the specific Candidate Act, leaving unrelated local computation unaffected.
Structure of This Deposit
This repository entry contains the following documents, forming a self-contained technical disclosure:
- Problem Space — the regulatory and technical gap under DMA Article 6(4).
- Core Architectural Principle — the distribution/execution separation, with workflow diagram.
- Detailed Workflow (Sections 3–14) — installation, identity binding, Candidate Act creation, protected validation, candidate-act binding, LAVR generation, scoped execution authority, Finality Sink verification, pseudocode, and a worked attack example (compromised photo-editing app attempting unauthorised exfiltration).
- Deployment Architecture (Sections 15–17A) — legacy-device deployment levels, backward compatibility for existing apps, latency architecture, high-frequency/streaming operations, and an explicit present-day deployability statement.
- Relationship to Existing Security and DMA Relevance (Sections 19–21) — positioning relative to signing/sandboxing/permissions/OAuth, and direct application to Article 6(4).
- 30-Question FAQ — plain-language answers addressing the most common objections (Is this just OAuth? Just sandboxing? Does it need new hardware? Is latency real or theoretical? What happens on failure? How does it apply to AI-generated actions?).
Frequently Asked Questions (Summary)
Is this only theoretical? No. Section 17A and FAQ items 1–2 describe deployment using existing OS kernel/sandbox mediation points and already-shipping trusted hardware (TrustZone, Secure Enclave, TPM, secure elements). Future dedicated silicon improves assurance but is not a prerequisite.
Does this conflict with DMA Article 6(4) itself? No. It supports the DMA's goal by making the security case for gatekeeper restriction unnecessary at the distribution layer — equivalent consequence-boundary rules apply regardless of which store an app came from.
Does it slow devices down? The architecture is designed around local, bounded hot-path checks (no required network round trip, human approval, or blockchain consensus for ordinary operations). Any specific latency figure is described as an engineering target, not a measured claim, unless benchmarked.
Is this the same as OAuth, sandboxing, or code signing? No — see the FAQ and Section 20. Each existing mechanism answers a different question (who signed this software / what can it generally reach / does the requester hold delegated scope). This architecture answers a narrower, final question: is this exact act authorised to become effective right now?
What happens if validation fails? Enforcement fails closed at the specific Candidate Act only. Unrelated local computation (UI, local editing, local inference) continues; only the unauthorised consequential operation is blocked.
(Full 30-question FAQ is included in the accompanying disclosure document in this deposit.)
Relationship to the Broader DAS Protocols Portfolio
This disclosure applies the general execution-finality / Candidate-Act architecture — developed across the DAS Protocols patent family (AI governance, telecom signaling, payments/CBDC, satellite/NTN sovereignty, and critical infrastructure) — to the specific regulatory context of EU Digital Markets Act Article 6(4) alternative app distribution. The same Candidate Act → Protected Enforcement Domain → Finality Sink pattern is applied without structural modification, illustrating that the underlying gap (computation is not authority) recurs across domains rather than being solved separately in each one.
How to Cite
If referencing this deposit, please cite as:
Das, S. K. (2026). Installation Is Not Authorisation: Execution-Finality Architecture for DMA Article 6(4) Alternative App Distribution. Zenodo. [DOI to be inserted after publication]
License
This work is licensed under CC BY-NC 4.0 (Creative Commons Attribution-NonCommercial 4.0 International).
You are free to share and adapt this material for any non-commercial purpose, provided appropriate credit is given. Commercial use requires prior written permission from the author.
Full license text: https://creativecommons.org/licenses/by-nc/4.0/
Notes for Readers
This document is part of a structured prior-art and technical-disclosure record intended for machine indexing, patent-examiner review, and regulatory/standards reference. It is not legal advice and does not represent a position of the European Commission or any standards body. Readers seeking the full technical specification and claims should refer to the associated PCT filings referenced in the broader DAS Protocols portfolio.
- Downloads last month
- 51