Workload Identity

Workload identity is the identity a piece of software — a service, container, function, or agent — presents to other software, derived from what and where the workload is rather than from a stored secret. It is how non-human actors authenticate to one another.

Why it matters

A shared secret in configuration is the classic way for services to authenticate, and it fails at scale: it is long-lived, copied everywhere, hard to rotate, and impossible to attribute to one instance. Workload identity replaces the shared secret with a short-lived, verifiable credential that the platform issues and rotates automatically. If the credential leaks, it expires; if a workload moves, its identity follows from its attested properties rather than from a file on disk.

The SPIFFE model

The best-known implementation is SPIFFE (Secure Production Identity Framework for Everyone), with SPIRE as its reference implementation. Its vocabulary:

  • A SPIFFE ID uniquely identifies a workload, namespaced within a trust domain.
  • An SVID (SPIFFE Verifiable Identity Document) is the document with which the workload proves its identity. It comes as an X.509 certificate (for mutual TLS) or as a JWT (JWT-SVID), whose `sub` claim carries the SPIFFE ID.
  • A trust domain bounds which authorities may issue identities a given system will accept.

Because SVIDs are short-lived and keys are rotated, the workload holds no static secret to steal.

The IETF work

The IETF is standardising the wider architecture. The WIMSE (Workload Identity in a Multi System Environment) working group’s architecture document defines a workload as software executing for a specific purpose, and describes protocols and payloads for conveying workload identity and security context across systems — including trust domains, workload identifiers, and cross-domain communication. This matters for agents, which frequently call across organisational boundaries where no shared platform attests them.

Workload identity also underpins zero trust: NIST SP 800-207 makes identity and per-request policy, rather than network location, the basis for access.

Workload identity vs agent identity

Workload identity answers what is this software? It does not, on its own, answer whose agent is it, and what is it mandated to do? A SPIFFE ID can tell a service that the caller is the expected binary in the expected trust domain; it cannot tell the service which principal that process is acting for or what it is authorised to do. Agent identity — and KYA — add that layer. See Why agent identity is a new problem and the standards tracker.

Related terms

Sources

  1. SPIFFE, SPIFFE Concepts (Accessed 2026-10-08)
  2. SPIFFE, JWT-SVID specification (Accessed 2026-10-08)
  3. IETF, draft-ietf-wimse-arch, WIMSE Architecture (Accessed 2026-10-08)
  4. IETF, WIMSE Working Group documents (Accessed 2026-10-08)
  5. NIST, SP 800-207, Zero Trust Architecture (Accessed 2026-10-08)

Know Your Agent (KYA) explains agent identity, verification and accountability. This is an explainer, not legal or compliance advice — see our Sources & methodology. New to KYA? Start with What is Know Your Agent? and the glossary.