Eggs. Meet basket.

Yesterday’s email concerning the Google outage was really an email about eggs and baskets. Which raises the obvious question: “Should I spread everything across many different vendors?”

Sort of, but not quite. There are two scenarios here where fixing one can cause the other:

The one basket mistake: everything going through a single provider with no fallback. The organisations hit hardest by the Google failure were the ones with no plan B, who’d never thought about what they’d do if that happened.

The hundred baskets mistake: An extra CDN, a bunch of plugins, another SaaS tool for authentication, etc. Every addition that feels like prudence is also another door added. Supply chain attacks love those doors. Why bother trying to batter down the front door when you can slip in through a side door like an Identity and Access Management provider. A sprawl of integrations doesn’t make you resilient, it just creates a larger attack surface.

The question to ask for any service you add is: if this disappears or turns on me, what happens?

With a single basket, the answer is grim but easy: “everything stops”.

With a tangle of baskets, the answer can be: “something I didn’t know was there breaks or, worse, becomes malicious”.

Build fallbacks, not labyrinths. Know what’s critical and what’s convenient. Test your exits. Prune regularly.

Colin