Observations & Insights

Infrastructure bake time

Sam Lessin posted an interesting observation that the speed of execution on new infrastructure innovations may actually create some of the technical debt long-term.

With fast execution, we roll out the first design and then pay for it long-term. The money has been spent and the infrastructure has been laid. We see a tangential version of this when emerging countries (who roll out infrastructure slightly later) get more advanced technology and are able to take advantage of more bake time.

What are some ways to offset this? Should set a pace (a bake time) for rolling out new infrastructure that takes this into account?

Type 1 vs Type 2 Infrastructure

One way to categorize innovations in order to estimate bake time would be to break out innovations into two types: 1) easily upgradeable and 2) nonupgradeable infrastructure.

Type 1: Nonupgradeable Infrastructure

With nonupgradeable infrastructure, the infrastructure would take extraordinary effort and cost to upgrade. This is the infrastructure where bake time matters a lot. And, a phased approach may be preferable.

Type 2: Upgradeable Infrastructure

Upgradeable infrastructure may be a design pattern goal for infrastructure. In other words, perhaps pay a little extra in order to create an infrastructure that you can upgrade easier. Also if the infrastructure is upgradeable then we can lessen the impact of bake time.

This is a spectrum, so we would want to figure out whether infrastructure is more like type 1 or 2.

The other thing to call out is that people often assume that speed is almost always good, especially when discussing progress. But, it isn’t. And, we need discussion and discovery around how to most sustainably create progress – to balance against the assumption of speed.

Leave a Reply

Your email address will not be published. Required fields are marked *