The official’s judgment
An authorization is a risk decision by a named person. It takes what it takes, and rushing it is how programs lose the second one.
Not a universal promise: a planning model for the case where a usable vehicle exists, the boundary is acquired rather than invented, long-lead connections start on day one, and named owners let engineering, assessment, evidence and decisions overlap. It is the complementary follow-on to readiness, taking the mission system through its own authorization on the evidence readiness already built.
The readiness program runs inside the landing zone’s own authorization—typically 24–28 weeks, and the provider and JWCC work stream rather than this one. This calendar counts from the task order and assumes the boundary can be acquired pre-engineered. Where the landing zone is engineered and authorized through the provider stream instead, that clock sets the pace, and the readiness work runs inside it. Either way, the authorizing official owns the decision.
They overlap—that overlap is where the calendar is won—but each has one owner and one artifact that says it is done.
Cloud is bought, not requisitioned. An enterprise vehicle—the Department’s joint cloud contract, or your agency’s own—turns a funded need into a task order that can pay a cloud provider this fiscal year.
Systems reach a government cloud region over a dedicated, approved network path, and circuits and connection approvals are the longest lead item nobody starts early. Ordering the path the week the task order signs is free schedule; ordering it in month four is a stall you chose.
This is the move that sets the calendar. A landing zone can be engineered from scratch, or acquired as a pre-engineered boundary with the security stack already wired and the paperwork already mapped to the controls.
A test authorization lets real testers use the system for a bounded window. An independent assessor examines the boundary and the package. The monitoring that started the day the boundary went live accumulates the operating evidence the decision will rest on—roughly ninety days of it.
An authorization to operate is a risk decision by a named authorizing official, not a certificate that falls out of a toolchain. Then come production users, live data, and continuous monitoring for as long as the system runs.
Moves 01–03 run in parallel from day one. Move 04 starts the day the boundary is live. Move 05 compresses for exactly one reason: everything handed to the official is finished, monitored, and familiar.
The longer path is not a slower team. It is the same team building in sequence what the eight-month path acquires: the landing zone designed and built, the security stack procured and tuned, the package written from a blank page, and only then the proving window and the decision.
A calendar this short earns skepticism. Here is each month-saving claim and what it rests on, because the compression is real, and it is also specific.
Hand-rolling a landing zone means designing the account structure, guardrails, logging, key management and network segmentation from first principles, then documenting all of it. A pre-engineered boundary arrives with those decisions made, deployed, and already described in authorization language, so the calendar spends weeks configuring it to the mission instead of quarters inventing it.
Vulnerability scanning, centralized logging, endpoint protection and alerting are in the boundary on day one rather than procured, integrated and tuned one tool at a time. That is also why the evidence clock can start at month three: there is a monitoring system running to produce evidence.
The security plan, policies and procedures for the boundary’s controls exist before the engagement starts, matured across prior authorizations. Your team writes the part only it can write—the mission system’s own behavior.
In an acquired boundary, most controls are implemented by the platform, shared with it, or inherited from the cloud provider’s own authorization. The mission team’s share is the slice only it can own; fewer controls to satisfy is fewer months to satisfy them.
Assessment stretches when the assessor meets a novel architecture and a hand-written package. A boundary pattern assessed before, described the same way each time, gets examined rather than reverse-engineered.
The most common stall is sequencing: award, then order the circuit, then start the boundary. Every lane that can start on day one does, and the long-lead network path is ordered in week one because nothing else on the chart can absorb its delay.
It compresses engineering and paperwork, and refuses to compress judgment, testing, or evidence. An eight-month plan that shrank those too would not deserve anyone’s signature.
An authorization is a risk decision by a named person. It takes what it takes, and rushing it is how programs lose the second one.
Roughly ninety days of the system actually operating, monitored. Artifacts can be accelerated; operating history can only be accumulated.
Real testers, bounded time, the same front door production will use. A skipped test authorization is a finding, not a shortcut.
The plan assumes your named owners show up and decide weekly. Every deferred decision is added to the end of the calendar, not saved.
A marketplace acceleration service can compress engineering and documentation, and continuous monitoring and authorization support can structure the evidence stream. Neither can grant or transfer an authorization. Everything on this page exists to put a complete, honest picture in front of the one person who decides—and anyone who implies otherwise is selling theater.