Business Continuity Architecture, Not a Binder

Business continuity architecture is the restore you can run, the login you can revoke, and the registrar you still hold when the vendor is gone. A binder on a shelf is not architecture. NIST SP 800-34 Rev. 1 treats testing and exercises as a step in the plan, not a cover page. If the last restore has no date, you have a story. Staff the architecture, or the Saturday will staff you.

I run Kief Studio with Brian. If we were walking your stack on a quiet afternoon, I would ask for the last successful restore, whose name is on the registrar, and who gets certificate expiry mail. Pretty homepages do not answer those. Architecture does.

What business continuity architecture actually is

NIST SP 800-34 Rev. 1 (May 2010, updated 11 November 2010) is a contingency planning guide for information systems. It is federal language. The useful part for a small operator is the sequence: policy, business impact analysis, preventive controls, recovery strategies, the plan, then testing, training, and exercises, then maintenance. Most shops stop at the plan. The plan is step five of seven.

ISO 22301 is the international requirements standard for a business continuity management system. ISO's brochure says it applies regardless of size, industry, or nature of business. The 2019 edition was amended in 2024. You do not need a certificate on the wall. You need the ability to keep delivering at a capacity you named in advance. That capacity is a number of hours and a named person, not a slogan.

Architecture here means the boxes, the backups, the identity, and the drill. Who can restore. Where the copy lives. How you prove the copy is not a screenshot. What is included in managed hosting is the hosting column of the same drawing.

Why binders fail on a Saturday

The binder lives in a folder nobody opened since the insurance questionnaire. The restore lives in one person's head. When that person is on a plane, the binder does not SSH. Tribal knowledge is the undocumented step. When they leave, the runbook leaves with them. Fragmentation wearing a continuity label.

I have sat with owners who could show me a 40-page continuity PDF and could not name the registrar. The backup was a folder on a laptop. The certificate mail went to a freelancer who had changed jobs. None of that felt like a crisis until a Saturday. Then it was the only conversation in the room.

Keep the registrar in your company email. A retained operator can hold DNS and deploy keys. They should not be the only name on the domain. Who owns the site, the host, and the login is the ownership list. Continuity is what you do with that list at 2am.

The drill is the architecture

Pick one workflow. Restore it on a quiet afternoon. Time it. Write the hours, the person, and the date on one page. That page is worth more than a binder. NIST calls this testing, training, and exercises. I call it a fire drill for the stack. A restore you have never run is a story.

Do it while everyone is calm. Do not wait for a price change or a terms update to discover you cannot leave. Own versus rent AI is the same drill for prompts and logs. Export is how you find out if you own the layer.

Business impact analysis, in NIST's sequence, is how you name what has to come back first. For a small operator that is often the site, email, and the payment form. Write those three in order. Restore in that order on the drill. If the payment form depends on a plugin nobody listed, you just found the architecture. Put the plugin on the page. Name who updates it. That is more continuity than a second appendix.

What to write down, in eight lines

Whose email and card sit on the registrar. Whose dashboard holds DNS and certificates, and who gets expiry mail. Where backups live, and the last restore date. Who holds CMS and deploy keys, and how those keys are revoked. What is the named capacity you can keep during a disruption: hours, not adjectives. Who is on the call at 2am. How you leave with a dump in a week. When you last walked the list.

If four of the first six are the last freelancer, you do not have continuity. You have a launch. Partner vetting is how you stop hiring that failure mode. One vendor versus four is why four half-runbooks cost more than one operator.

At Kief Studio the public rule is the same as on ltfi.ai. You keep the registrar, the words, and the customer data. We hold the operational login so the night shift is staffed. Clients do not configure the fleet. kief.dev and briansgagne.com are the engineering notes. I write the drill. He designs the box the drill runs on.

Related reading

• What is included in managed hosting
• Who owns the site, the host, and the login
• What self-hosting actually costs
• Data governance, where to start
• Cost savings from consolidating compliance audits
• What LTFI is, and what you do not install

Frequently Asked Questions

What is business continuity architecture for a small company?

The restore you can run, the logins you can revoke, and the registrar you still hold. It is people and systems, not a PDF on a shelf.

Do I need ISO 22301 certification?

Not to start. ISO 22301:2019 is a requirements standard for any size of organization. Use it as a checklist. A dated restore beats a certificate you cannot staff.

How often should we test a restore?

NIST treats testing as an ongoing step. For a small operator, a quarterly restore of one real workflow is a serious start. Write the date.

Is a daily backup enough?

No. A backup without a restore date is a story. The test is whether you can bring the site back, and who can do it.

Where does this sit next to managed hosting?

Managed hosting is who does the night shift. Continuity architecture is whether that night shift can actually restore. Ask for both.