End-to-end verifiability proves the count. Attestation shows which software the servers ran while they counted.
An electoral authority offers internet voting. The protocol is end-to-end verifiable: each voter can check that their vote was recorded, and anyone can check that all recorded votes were counted. The mathematics is public and has been reviewed. The public still asks a plainer question: what was running on those servers on election day?
The software was audited months before the vote. Since then it was rebuilt, patched and deployed to several sites. Verifiability protects the result even if a server misbehaves, as long as at least one of the independent components is honest. That condition is an assumption about machines. An audit report from spring does not show what a machine ran in autumn.
The audited source is built reproducibly, and the build's measurements are published before the vote. Each server measures what it boots and reports it to verifiers run by different parties, including independent observers. Measurements are checked at the start, at intervals during the voting period and before the count. The results go into a public log. A server that fails is taken out, and the protocol's redundancy covers it.
The cryptography still proves the count. Attestation adds evidence for the assumption underneath it, in a form observers can check themselves. It turns 'the audited software was installed' into something that can be shown for every hour of the vote.
An illustrative case built from public documentation and common architectures. It describes no particular company.
Working on attested cloud workloads? A first call takes 30 minutes and commits you to nothing. Bring your product and the one question that worries you.
Next fixed date: EU Digital Identity Wallets are due in every member state on 24 December 2026, 42 days from now.