Skip to content

Cryptographic inventory, the first step nobody wants to take

A cryptographic inventory records every place cryptography is used: algorithms, key sizes, certificate lifetimes, libraries, protocol versions and the hardware that executes them. It is the first migration step, the least popular, and the one that determines whether the rest of the programme is planned or discovered.

Why it is skipped

It produces no visible improvement, takes months, and the output is a spreadsheet. Every incentive points toward starting with something demonstrable instead.

The consequence of skipping it is that the migration finds its own scope, in production, at the least convenient moments.

What to record

  • Algorithms and key sizes in use, per system and per protocol.
  • Certificate inventory with expiry and renewal path, including internal PKI nobody has looked at recently.
  • Libraries and versions — the actual implementations, since that is what has to be upgraded.
  • Protocol versions and negotiated parameters, which are frequently not what the configuration says.
  • Hardware dependencies: HSMs, TPMs, secure elements, network appliances, smartcards.
  • Data confidentiality horizons, so harvesting exposure can be assessed where it genuinely applies.

The part most inventories miss

Hardware. Software inventory is tractable with scanners and dependency analysis. The cryptography executing inside an HSM, a network appliance or a secure element is not visible to any of those tools, and it is the part that cannot be upgraded by a package manager.

This is where migration timelines are actually set. A library upgrade is a release; an appliance that cannot support new algorithms is a procurement cycle and a qualification effort. Find those early or they will find you.

The questions to ask of each hardware item

  • Can it support post-quantum algorithms at all, and in what form — firmware update, or replacement?
  • If firmware, is the update available now, promised, or unquantified?
  • What is the requalification cost after that update?
  • When does this device reach end of support, relative to the timeline that applies to you?
  • Where does it get its entropy, and can that be evidenced? The migration is a rare moment when someone is looking at these systems at all — and the audit questions are worth asking while the lid is off.

The useful by-product

Most organisations discover cryptography they did not know was running, certificates nobody owns, and appliances past end of support. That is worth the exercise independently of any quantum computer.