ScrutoraCode, cloud & consent
Back to Blog
|6 min readCompliance Guide

DPDP Rules 2025, Rule 6(1)(b): What “Access Control on Computer Resources” Actually Requires

Part 2 of 7. “Computer resource” borrows the IT Act’s definition, and it’s broader than most access programmes assume.

By Vimalendukumar Dwivedi, Co-founder - Scrutora

Part of a series on DPDPA Section 8(5) and DPDP Rules 2025, Rule 6. Read the pillar post first if you haven’t.

The Clause, Verbatim

“(b) appropriate measures to control access to the computer resources used by such Data Fiduciary or such a Data Processor, wherever applicable;”

DPDP Rules 2025, Rule 6(1)(b)

Rule 6(2) adds one more piece: “computer resource” carries the same meaning given to it in the Information Technology Act, 2000. That’s a deliberately broad, technology-neutral definition, and it’s worth reading in full before assuming this clause only means “who has database access.”

What It Technically Requires

Under the IT Act’s definition, “computer resource” covers computer systems, networks, data, computer databases, and software, essentially the full stack a Data Fiduciary or Processor uses to touch personal data, not just the data store itself. Read that way, Rule 6(1)(b) requires access control across:

Application-layer access. Role-based access control (RBAC) governing which users and services can read or write personal data through your product’s own interfaces and APIs, with roles scoped to what each function actually needs rather than broad admin defaults.

Data-layer access. Database-level permissions independent of application logic, so a compromised application credential doesn’t automatically mean unrestricted database access. This includes row-level or column-level restrictions where a single table mixes personal data with non-personal data.

Infrastructure access. Who can SSH into servers, access cloud consoles, or modify infrastructure-as-code that provisions systems processing personal data. A well-secured application sitting on a cloud account with broad, unaudited IAM permissions has a real gap here, even if the application layer looks clean.

Authentication strength. Multi-factor authentication for any account with access to personal data, particularly privileged or administrative accounts, and particularly for Data Processors who may be accessing your systems remotely.

Offboarding and review. A process for revoking access when someone changes roles or leaves, and periodic review of who currently has access versus who should. Access granted correctly six months ago and never revisited is a common gap between “we have access control” and “our access control reflects reality.”

Where This Gets Subjective

Granularity. “Control access to computer resources” is satisfied, minimally, by an application login screen. It’s satisfied more thoroughly by RBAC scoped to specific data fields, and more thoroughly still by attribute-based access control that factors in context (device, location, time) before granting access. The clause doesn’t specify a granularity, so organisations reasonably disagree about how far down the stack “control” needs to reach before it’s genuinely controlling anything.

What counts as a “computer resource” in your environment. Does the definition extend to developer laptops that have local copies of production-adjacent data? CI/CD runners that process personal data during test runs? A contractor’s personal device used to access an admin panel? The IT Act’s definition is broad enough to plausibly capture all of these, but organisational access-control programmes very often stop at production infrastructure and treat everything upstream of it (developer environments, build pipelines, third-party tooling) as out of scope. That’s a defensible operational boundary. It’s a much harder one to defend as a legal boundary if a breach originates from exactly that upstream environment.

“Wherever applicable.” The clause itself carries a built-in qualifier: access control is required wherever applicable, which implicitly concedes that not every computer resource needs the same treatment. Reasonable people will disagree on where “applicable” starts and stops, particularly for lower-sensitivity data or internal tooling several steps removed from the personal data itself.

What the Breakdown Really Means

Rule 6(1)(b) is easy to satisfy shallowly (a login page exists) and hard to satisfy thoroughly (access is scoped, authenticated strongly, reviewed periodically, and extended consistently across every layer of the stack that can reach personal data, not just the layer that’s easiest to audit). The clause’s breadth, inherited from the IT Act’s definition of computer resource, is the part most likely to be under-scoped in practice.

If your access-control programme starts at the production database and stops there, Rule 6(1)(b) is asking you to look one layer further up the chain than that.

Vimalendukumar Dwivedi is Co-founder of Scrutora, a Code Compliance Platform.

See who can actually reach your personal-data endpoints

Scrutora flags missing or overbroad access controls across your application, database, and infrastructure layers, mapped to the specific DPDPA clause each gap violates.

Share this article