The divergence is not anecdotal. It is measurable.
A 2025 study led by independent researcher Karan Saini, titled Poisoned Wells, tested DNS-level censorship across six major Indian ISPs, scanning roughly 294 million domains-effectively mapping the visible global domain name space. The findings are stark: of 43,083 domains identified as blocked, only 1,414 were blocked uniformly across all six ISPs.
In other words, over 96% of blocked domains are enforced inconsistently.
This is not because ISPs receive different instructions. They receive the same blocking orders. But they are free to decide how to implement them.
That discretion is the system’s defining feature.
That discretion is not just legal-it is economic.
Internal network cost modelling by large Indian ISPs, as cited in 2024 consultations submitted to the Telecom Regulatory Authority of India on traffic management frameworks, shows that DNS-level blocking costs less than ₹0.05 per 10,000 queries, whereas SNI-based filtering and deep packet inspection can increase per-query processing costs by 8–15x, depending on traffic load and encryption penetration.
Data published by Sandvine in its Global Internet Phenomena Report 2024 indicates that over 90% of Indian web traffic is now HTTPS-encrypted, sharply reducing the effectiveness of legacy HTTP interception and pushing ISPs toward either DNS-layer shortcuts or significantly higher-cost inspection systems.
This creates a rational but uneven equilibrium:
- Low-cost DNS blocking becomes the default for bulk categories like piracy and gambling
- High-cost filtering is reserved for politically sensitive or security-critical domains
The result is not merely inconsistent enforcement-it is cost-optimised enforcement, where uniformity is sacrificed for operational efficiency.
The Mechanics of Control: Cheap, Effective, Uneven
In practice, most Indian ISPs rely on the lowest-cost intervention layer: DNS manipulation.
When a user types a website address, DNS translates that name into an IP address. By altering this lookup-what engineers call DNS poisoning-ISPs can redirect or nullify access without ever touching the underlying content. It is inexpensive, scalable, and, according to measurement studies by the Open Observatory of Network Interference (OONI), widely deployed across jurisdictions with decentralised censorship models.
More sophisticated techniques-such as deep packet inspection or TLS-level filtering via Server Name Indication (SNI)-exist, but they are costlier and computationally intensive. As a result, they are used selectively, often for higher-priority categories such as national security or politically sensitive content.
This layered approach produces predictable asymmetry.
A piracy website might be blocked on one ISP and accessible on another. A politically sensitive domain may be uniformly blocked across all networks within hours. The system does not treat all categories equally, and it does not enforce them consistently.
What Gets Prioritised-and Why
The data reveals a hierarchy of enforcement.
Domains linked to terrorism, militancy, or geopolitical sensitivity show near-perfect consistency across ISPs. For instance, platforms such as Weibo or certain Kashmir-based publications have been uniformly restricted, reflecting coordinated, high-priority enforcement.
By contrast, categories such as copyright infringement, gambling, pornography, and peer-to-peer file sharing exhibit wide variance. Enforcement here appears driven less by urgency and more by ISP interpretation and operational convenience.
This selective consistency is not accidental. It reflects an implicit triage system.
High-risk categories are tightly synchronised. Lower-risk categories are loosely enforced.
The outcome is a censorship regime that is not binary (blocked vs accessible), but probabilistic.
The same system that produces fragmentation can, under specific conditions, deliver near-total uniformity.
That was visible in June 2020, when the Government of India ordered the blocking of 59 Chinese-linked applications, including TikTok, under Section 69A of the Information Technology Act, 2000, citing national security concerns.
Within hours of the order-based on telecom operator disclosures and network measurements tracked by Internet Freedom Foundation and global observability platforms like NetBlocks-access to TikTok and other listed apps dropped to near-zero across all major Indian ISPs and mobile networks.
The coordination was not incidental. It was engineered.
Unlike routine domain blocking, the TikTok order triggered multi-layered enforcement:
- App-store delisting: Google and Apple removed TikTok from their Indian storefronts within hours
- DNS and IP-level blocking: ISPs blocked both domain resolution and direct IP access
- CDN-level disruption: Traffic routing to TikTok’s content delivery endpoints was actively curtailed
- Platform compliance: ByteDance itself disabled certain services for Indian users
Data from Sensor Tower shows that TikTok had over 200 million users in India prior to the ban, making it one of the largest single-country user bases globally. Post-ban, usage effectively collapsed to negligible levels within days-a level of enforcement consistency rarely seen in India’s otherwise uneven censorship landscape.
What explains the difference?
Three conditions aligned simultaneously:
- Clear national security framing - reducing ambiguity in enforcement priority
- Finite, identifiable targets - a closed list of apps, rather than open-ended domains
- Centralised signalling + platform cooperation - ensuring alignment across ISPs, app stores, and the platform itself
In effect, the system moved from decentralised interpretation to centralised execution with ecosystem-wide compliance.
The contrast with routine website blocking is instructive.
Where targets are diffuse (millions of domains), enforcement is uneven.
Where targets are finite and politically prioritised, enforcement becomes near-total.
This reveals a deeper truth about India’s censorship architecture:
it is not incapable of consistency-it is selectively capable of it.
The Opacity Problem: A System That Cannot Be Seen Cannot Be Challenged
If fragmentation is one axis of the problem, opacity is the other.
Blocking orders issued under Section 69A are confidential by design. ISPs are legally bound not to disclose them. There is no central, publicly accessible registry of blocked domains. Users discover blocks only when access fails. Website operators often receive no prior notice.
This creates a structural asymmetry of information.
The safeguards envisaged in Shreya Singhal-particularly the right to be heard and the existence of a review mechanism-depend on visibility. Without disclosure, they become largely theoretical.
The gap between legal design and operational reality is now well documented.
Empirical analysis of blocking orders reviewed by the Software Freedom Law Center India shows that only a small fraction of affected website operators are ever formally notified, despite procedural requirements under the blocking rules framed pursuant to Section 69A.
Further, research compiled in 2023–24 by the Internet Freedom Foundation indicates that review committee proceedings-mandated to periodically assess blocking orders-are rarely disclosed and almost never result in publicly known reversals, limiting their practical corrective function.
This produces a structural asymmetry:
- The State retains full visibility over what is blocked
- Users and platforms operate in informational darkness
In such a system, procedural safeguards exist on paper-but cannot be meaningfully exercised without visibility of the underlying action.
Research by organisations such as the Centre for Internet and Society has repeatedly flagged this gap, noting that India’s blocking regime combines strong executive power with weak transparency obligations, limiting both judicial oversight and public accountability.
When Compliance Becomes Interpretation
The decentralised nature of enforcement introduces a second-order effect: ISPs do not merely implement orders-they interpret them.
This produces three recurring distortions:
- Overblocking: ISPs may block entire domains rather than specific URLs to minimise compliance risk, especially in copyright cases.
- Under-enforcement: Some ISPs fail to fully implement orders, either due to technical limitations or cost considerations.
- Persistence failures: Domains that have been officially unblocked remain inaccessible on certain networks, suggesting weak rollback protocols.
In effect, ISPs function as de facto regulators at the implementation layer-without the accountability or scrutiny that formal regulators face.
The Scale Effect: Why This Matters More in India Than Elsewhere
India’s internet user base-estimated at over 850 million users, according to joint 2024 estimates by the Telecom Regulatory Authority of India and the Internet and Mobile Association of India-amplifies the consequences of this system.
At this scale, even small inconsistencies translate into large systemic effects:
- A domain blocked inconsistently can still attract millions of users through alternate ISPs.
- A wrongly blocked domain can remain inaccessible to large user segments without recourse.
- Differential access can distort digital markets, especially for smaller platforms that rely on uniform discoverability.
For startups and digital businesses, this introduces a new variable: network-dependent visibility risk.
A service may not fail because of demand-but because it is unevenly reachable.
That risk is no longer hypothetical. It has already played out-quietly, and at scale.
In mid-2024, developers across India began reporting sudden outages while accessing backend services hosted on Supabase, an open-source alternative to Firebase used by thousands of early-stage Indian startups for authentication, databases, and APIs. The issue did not originate from Supabase’s infrastructure. Its global status dashboards showed normal uptime, and traffic from other geographies remained unaffected.
The failure was local-and network-specific.
Developers on networks operated by Reliance Jio and Bharti Airtel reported connection timeouts, while the same services remained accessible via smaller ISPs or through VPN rerouting. Packet-level diagnostics shared in developer threads archived on GitHub and corroborated by network telemetry tools from Cloudflare suggested DNS-layer interference and SNI filtering patterns consistent with domain-level blocking.
What made the incident particularly disruptive was its architecture-level impact.
Supabase is not a consumer-facing website. It sits behind applications-powering login flows, database queries, and real-time updates. When access to its domains failed, entire applications broke. Startups reported authentication failures, stalled transactions, and cascading backend errors.
According to incident reconstructions shared by Indian developer communities and cross-referenced with Cloudflare Radar traffic anomalies, some applications saw traffic drops of 20–40% within hours, not because users left, but because core services became unreachable on specific networks.
The resolution timeline exposed another structural gap.
There was no formal public explanation from ISPs, no disclosed blocking order, and no standard escalation channel. Restoration occurred unevenly over several days, suggesting either silent rollback of filtering rules or network-level reconfiguration-neither of which was publicly documented.
For founders, the lesson was immediate and operational:
- Redundancy is no longer just about cloud regions-it is about network-layer resilience
- Monitoring must extend beyond uptime to include ISP-specific accessibility checks
- Infrastructure choices now carry a hidden variable: regulatory exposure at the network level
In effect, a compliance mechanism designed for content control had transiently disrupted core application infrastructure-blurring the line between censorship and systemic fragility.
For digital businesses, this introduces a risk that rarely appears in pitch decks or regulatory filings: network-contingent discoverability.
Executives at Indian SaaS and developer infrastructure firms privately acknowledge that unexplained access disruptions-such as the temporary blocking of backend platforms like Supabase in 2024-can trigger double-digit traffic drops within hours, not because of user churn, but because of ISP-level inaccessibility.
According to internal incident analyses shared by developers on forums tracked by Cloudflare and GitHub communities, such disruptions often go undiagnosed for extended periods, as teams initially attribute them to server outages or application errors rather than network-level filtering.
The asymmetry is particularly severe for early-stage platforms:
- Large platforms (Google, Meta, Amazon) benefit from multi-layer redundancy and direct ISP coordination
- Smaller services rely on standard DNS resolution paths, making them more vulnerable to silent blocking
In effect, censorship infrastructure begins to shape competitive outcomes, not just content access.
A Hybrid Model with Structural Limits
Globally, censorship architectures tend to fall into two broad models:
- Centralised systems (e.g., China), where enforcement is uniform but tightly controlled.
- Platform-led moderation systems (e.g., the US/EU), where decisions are decentralised but increasingly transparent and contestable.
India’s model is a hybrid.
Authority is centralised. Execution is decentralised. Transparency is minimal.
What makes this architecture particularly durable is that it sits between regulatory silos.
India’s telecom regulation-anchored in licensing conditions enforced by the Department of Telecommunications-focuses on compliance with lawful orders, not on standardising technical implementation. Meanwhile, digital governance frameworks under MeitY focus on intermediary liability and content moderation, not network-layer enforcement.
As a result, no single regulatory body currently audits how blocking is implemented across ISPs.
Industry submissions to TRAI’s 2023 consultation on net neutrality enforcement flagged this gap explicitly, noting that absence of standardised blocking protocols can produce differential user experiences even under identical legal mandates-a condition that sits uneasily with the principle of non-discriminatory access.
The system, therefore, persists not despite regulatory oversight-but partly because it falls between regulatory mandates.
This combination creates a system that is legally robust but operationally uneven-and therefore difficult to audit, challenge, or standardise.
This unevenness is beginning to produce early signs of what network engineers informally describe as “access arbitrage.”
Traffic-routing data and latency optimisation strategies observed in reports by Akamai Technologies suggest that users-especially in urban India-are increasingly able to bypass ISP-specific restrictions simply by switching networks, using alternate DNS resolvers, or routing traffic through VPN overlays.
While this restores access for technically aware users, it introduces a two-tier system:
- High-agency users (with technical awareness) experience a near-open internet
- Low-agency users remain subject to ISP-defined access limits
Over time, this creates a divergence not just in access-but in information asymmetry across user classes, a phenomenon typically associated with tightly controlled information regimes rather than open digital economies.
The Deeper Shift: From Blocking to Filtering
What emerges is a conceptual shift in how censorship operates.
Traditional censorship is visible. A site is blocked, and the user knows it.
India’s system is increasingly invisible. Access varies quietly, inconsistently, and without explanation.
This transforms censorship from an act of prohibition into a system of filtration.
The difference is subtle but consequential.
Prohibition invites scrutiny. Filtration often goes unnoticed.
Why This Matters
The stakes are not limited to free speech. They extend to the integrity of the internet as infrastructure.
When access depends on ISP choice:
- Equality of access breaks down, undermining the idea of the internet as a common public resource.
- Legal certainty weakens, as identical orders produce different outcomes.
- Accountability erodes, because neither the state nor ISPs can be easily scrutinised.
For a country positioning itself as a global digital economy-through initiatives spanning digital public infrastructure, data governance, and AI deployment-this inconsistency introduces friction at the foundation.
The issue is not whether India should regulate the internet. Every major jurisdiction does.
The issue is whether a system that produces multiple, opaque versions of reality can sustain trust at scale.
Because once the internet stops being uniform, it stops being predictable.
And once it stops being predictable, it stops being neutral infrastructure-and starts becoming something else entirely.
Politics
Economy
Business
World
Defence
Law & Justice
Science & Technology
Rida Prime
About us