eMudhra to exit global SSL/TLS certificate business

Original Article ↗ Hacker News Discussion ↗

An Indian certificate authority, eMudhra, is exiting the global SSL/TLS business after researchers discovered that some of its CA private keys were trivially weak values such as the number 4, making it possible for anyone to forge certificates. Commenters argue that this is an unforgivable failure for a company whose core product is “digital trust,” and note the massive global engineering effort now required to revoke and replace affected chains across browsers, operating systems and embedded systems. The incident also prompts broader criticism of CA vetting, the Web PKI trust model, and regulatory and cultural factors around security practices in India and elsewhere.

Incident and Exit from SSL/TLS Business

  • eMudhra announced it will leave the global SSL/TLS CA market, framing it as a strategic shift because that line is only ~1% of revenue.
  • Many commenters see this as “you can’t fire me, I quit”: exiting because no browser or OS would ever trust them again, not as a voluntary pivot.
  • Some argue regulators in India have opened an investigation, but others doubt there will be serious local consequences.

Technical Failure: “4” as a Private Key

  • Core issue: several EC CA private keys were trivially small integers (e.g., 4, 56, 115), recoverable in milliseconds from public certs.
  • Commenters explain this is astronomically unlikely under proper key generation; consensus is that a key-generation/wrapping script was badly broken (e.g., only using one byte, or mixing up test vs real keys).
  • eMudhra’s explanation (script bug during HSM key wrapping, human error choosing wrong file) is widely seen as evidence of deep process and engineering failures, not a one-off slip.

Reaction to Press Release and Trustworthiness

  • The press release is heavily criticized as evasive and defensive: blaming “social media amplification,” stressing that “infrastructure was not compromised,” and emphasizing other business lines are “unaffected.”
  • Many see that as minimizing a systemic failure in a company whose entire brand is “digital trust.”
  • Multiple commenters say they would not trust any of eMudhra’s remaining PKI, identity, or OTP products after this; one notes their OTP implementation also appears weak.

CA Ecosystem, Governance, and Responsibility

  • Strong criticism extends to the CA/Browser ecosystem:
    • Root programs (Mozilla, Google, Apple, Microsoft) and CAB Forum processes are faulted for admitting and cross-signing a clearly immature CA.
    • Sectigo’s cross-signing of eMudhra is questioned; some argue cross-signers share responsibility.
    • Others note that approval relies heavily on third‑party audits rather than real operational scrutiny.

Revocation, Mitigation, and Platform Behavior

  • Chrome and Firefox are praised for fast revocation; Apple’s slower and somewhat opaque reaction on iOS/macOS draws concern, with people testing devices against public check sites and sometimes seeing inconsistent results.
  • Many discover how fragmented trust handling is: languages and runtimes (Python, Node, Java, Go), OS stores, and applications often ship their own CA bundles, forcing many separate updates.
  • Some argue any stack that cannot quickly distrust a CA is itself negligent; others highlight the enormous global engineering cost of mass CA revocations.

Alternatives and Reform Proposals

  • Recurrent suggestions:
    • Constrain CAs by geography or namespace (e.g., country CAs only for their ccTLDs).
    • Use DNSSEC/DANE or similar mechanisms to bind keys directly in DNS, reducing CA power.
    • Make certificate pinning or stronger CAA/DNS-based constraints more common.
  • Some argue for a much smaller set of non-profit or heavily scrutinized CAs (e.g., “Let’s Encrypt–like”), while others note even well-run CAs have had issues; no clear replacement gains consensus.

Cultural, National, and Organizational Themes

  • Many comments blame Indian corporate and regulatory culture (e.g., “face-saving,” “izzat,” cronyism, low pay, and box‑ticking compliance).
  • Others push back, arguing similar corporate denial and regulatory failure happen worldwide, and that the real issue is weak institutions, audits, and incentives rather than nationality alone.
  • Several note that eMudhra’s leadership appears non-technical, with low developer salaries cited as symptomatic of underinvestment in competence.

Broader Conclusions

  • Widespread sense that:
    • eMudhra’s CA business is effectively dead, whatever the spin.
    • This incident is a “hall-of-fame” PKI failure that will drive tighter requirements (e.g., attested HSM keygen, better ceremony automation, more aggressive monitoring).
    • The Web PKI model remains fragile: any trusted CA’s catastrophic error or compromise imposes huge global costs, and this event has renewed calls to rethink or at least significantly harden that trust model.