GAL-2 Technologies

08/30/2026

72 hours. Two platforms. Both PASS.

GAL-2 Node RC11 completed real-duration upstream-outage endurance runs on macOS Apple Silicon and Linux ARM64.

LIVE → HOLDOVER → DEGRADED → FAIL_CLOSED → REJOIN → LIVE

No backward steps observed.

Public evidence:
DOI 10.5281/zenodo.22164457

Next step: independent reproduction.

DM for a temporary evaluator API key.

08/21/2026

The timing stack has a missing layer.

NTP, PTP, GNSS, atomic clocks and timing appliances are built to deliver time.

GAL-2 adds something different: application-facing time governance.

It governs whether an enrolled application should consume time before that time becomes trusted state.

We recently exercised GAL-2 under Telstra-class failure conditions involving upstream loss and recovery.

Today, the public RC11 macOS artifact is running a real-duration 72-hour continuous upstream outage test.

Your timing stack delivers time.
GAL-2 governs its consumption.

A new infrastructure category is emerging at the application boundary.

TimeGovernance

08/16/2026

Most systems try to make the clock better.

GAL-2 asks a different question:

Should the application trust time right now?

If the timing stack degrades, should the app keep going or know when to refuse?

08/15/2026

GAL-2 Node RC10 is now open for technical evaluation on Linux ARM64.

GAL-2 API creates GAL-2 Time.
GAL-2 Node makes it locally consumable.
The Time Contract governs its use.

What should an application do when time can no longer be safely consumed?

RC10 demonstrates bounded HOLDOVER, controlled recovery and REJOIN, FAIL_CLOSED behavior, and no silent fallback to raw host time.

Keep your timing stack.
Install the GAL-2 Node.
Connect the application once.

Technical evaluation and early pilot access are now open.

Need an evaluation API key? DM us or email [email protected].

08/12/2026

Something new is coming to the timing stack.

For years, applications have been expected to consume whatever time the system gives them. GAL-2 takes a different approach.

The GAL-2 API creates GAL-2 Time.
The GAL-2 Node makes it locally consumable.
The Time Contract governs its use.

The Node sits alongside your existing timing infrastructure. It does not replace NTP, PTP, or GNSS, and it does not discipline the host clock.

Keep your timing stack.
Install the GAL-2 Node.
Connect the application once.

Internal validation and endurance testing for the GAL-2 Node v1.0 release candidate are complete, including bounded holdover, controlled rejoin, fail-closed behavior, and application-level consumption without silent fallback to raw host time.

Technical evaluation access opens in September 2026 at gal-2.com.

GAL-2 is not another clock.

It’s a new application-level layer for deciding when time is safe to consume.

More soon.

07/20/2026

On July 6, we presented GAL-2’s application-boundary time governance model to the Open Compute Project.

Two days later, a national telecom outage showed what happens when a system accepts an invalid date as operational truth.
A server believed it was 2006.
Downstream systems believed it too.

That is the gap GAL-2 Node is being built to address.

The problem is not only when time fails.

It is when the system believes it.



https://www.theguardian.com/business/2026/jul/17/telstra-missing-software-update-undocumented-design-change-outage

07/18/2026

Coming next: GAL-2 Node.

A local delivery and governance layer designed to provide:

• GAL-2 Time as the default time path for enrolled applications
• controlled rejoin and explicit uncertainty at the point of consumption
• governed holdover and fail-closed behavior when time can no longer be justified

The existing UTC, GNSS, PTP and NTP stack remains in place.

GAL-2 Node does not replace or discipline the host clock. It creates a parallel, opt-in path for applications that need to know not only what time it is, but whether that time is still safe to consume.

Protected Core. Local governance. Application-level control.

Evolving from the daemon architecture already exercised through public experiments.

Development is already underway.

07/11/2026

I did not start GAL-2 by trying to build another clock.

I started with a question:

Why does software treat time as perfectly reliable when the systems delivering it can degrade, drift, correct, or fail?

That question became Fractal Time research, and eventually GAL-2.

UTC remains the reference. GAL-2 does not replace it.

GAL-2 adds a governance layer between delivered time and application state, helping software determine whether time is safe to consume before it becomes a payment, log, transaction, or permanent record.

I built it as a solo developer in Puerto Rico.

The evidence is public:

508,548 observed samples over seven days with zero backward steps.

A 120-hour characterization with FAIL_CLOSED at the declared 72-hour policy boundary.

Protocols declared and hashed before ex*****on.

Permanent public DOIs.

GAL-2 is not another NTP, PTP, or grandmaster.

It is governed time for application state.

When your time source quietly degrades, does your software know?

- F. Torres

gal-2.com/validation

07/08/2026

Most systems are designed to receive time.
Far fewer are designed to decide whether that time is still safe to use.

GAL-2 places a policy boundary between time sources and application state, helping software validate, evaluate, and authorize time before it becomes permanent state.
Receiving time is not the same as safely consuming it.

Govern time before time becomes state.

06/26/2026