When DNS Refuses to Shake Hands | Alt Production Group

What fighting BT taught us about infrastructure, interoperability and why businesses shouldn't have to fight their connectivity to innovate

There is a particular kind of infrastructure problem that, when you are neurodivergent, eventually stops existing in your head as logs, packets and configuration files and becomes a very literal mental image.

For us, DNS resolution with BT became two people trying to shake hands.

One person puts their hand out.

The other person slowly starts putting their hand out.

By the time their hand gets there, the first person has decided nobody is shaking their hand and pulled it back.

The second person eventually arrives with their hand out.

Nobody is there.

So they pull it back.

Then the first person tries again.

Repeat.

Forever.

Welcome to infrastructure.

Back in April, Alt Production Group found itself fighting through DNS resolution and connectivity problems while trying to operate infrastructure over a BT/Openreach connection.

To be clear, we haven't somehow escaped Openreach since. Openreach remains part of the connectivity underneath us. Nor is this an article claiming that BT's security is inherently bad. Security controls exist for good reasons.

Our frustration was with what happens when a provider's assumptions about how its customers should use a connection collide with the requirements of a business doing something less conventional.

BT has its own security architecture, its own controls and its own view of what constitutes an appropriate configuration. From the provider's perspective, that can make perfect sense.

From our side, however, we were trying to build and operate our own infrastructure.

And those two worlds did not always shake hands particularly elegantly.

Security can become a constraint

There is an uncomfortable question buried inside this.

At what point does security designed to protect a customer begin constraining a customer who already understands the security architecture they want to operate?

The same applies to commercial segmentation.

We understand that telecommunications providers are businesses. Different capabilities are packaged into different products. Enterprise functionality costs money. Revenue matters.

We are a business too. We rather like revenue.

But infrastructure is also an enabling layer for every other business sitting on top of it.

When capabilities necessary for experimentation, self-hosting or unconventional infrastructure become difficult to access, the consequence isn't simply inconvenience. You can create another barrier between somebody having an idea and somebody actually being able to build it.

For a small business or solo founder, those barriers compound quickly.

It is one of the experiences that has shaped how we think about infrastructure across Alt Production Group.

Infrastructure should enable capability

We don't think infrastructure should force every organisation into exactly the same operating model.

Security matters.

Reliability matters.

Commercial sustainability matters.

But so do interoperability, understandable controls and the ability of technically capable organisations to decide how their own systems operate.

That doesn't mean removing safeguards.

It means asking whether safeguards are actually protecting somebody or simply preventing them from doing something the architecture didn't expect.

There is a difference.

Alt Production Group increasingly approaches infrastructure from that position: give organisations strong defaults, strong protections and clear boundaries, but preserve meaningful control.

Because sometimes the technical problem isn't that either side is incapable of shaking hands.

Sometimes they just fundamentally disagree about when the handshake is supposed to happen.

And somewhere in Manchester, an AuDHD founder is staring at a terminal imagining two very confused hands repeatedly missing one another.

That is apparently how infrastructure companies get started.