N100FN100F
Published on

Multicloud and hybrid, with judgement

Using several clouds or combining cloud with your own infrastructure can be the right decision or a source of complexity with no return. When it pays off, when it does not and which pattern to choose.

Service line Cloud Architecture

A fashion with a cost

'Multicloud' is often presented as a virtue in itself: not depending on one provider. But every additional provider doubles the knowledge, the tooling, the security models, the billing formats and the points of failure. Without a concrete reason, multicloud is complexity you pay for every month.

The same goes for hybrid: keeping your own infrastructure connected to the cloud makes sense in some cases and is mere inertia in others.

Reasons that do justify it

More than one cloud?A decision tree: with no concrete reason, one provider; with a reason, the pattern follows the reason.ARCHITECTURE · MULTICLOUD AND HYBRIDMore than one cloud? It depends whyIs there a concrete reason?NoYesA single providerSimpler and cheaperWhat is the reason?Regulation orlocal dataConnected hybridThe best servicefrom each cloudPrimary provider+ specific workloadsBeing able to moveworkloadsPortabilityby designContinuity againsta provider outageActive-active(the most expensive)IN EVERY CASEUnified identityPrivate network and zonesCommon governanceCost in one formatN100F · n100f.com

  • Regulation or residency: data that has to stay in one place or on your own infrastructure.
  • Latency or local dependency: industrial systems, branch offices or equipment that cannot depend on the internet.
  • The best service from each provider: a specific capability (analytics, AI, a managed service) that justifies operating a second cloud for that workload.
  • Continuity in the face of a provider outage: a real requirement, and an expensive one, that has to be designed for from the start.

'So as not to depend on anybody' or 'to negotiate better' rarely justify the cost of operating two platforms on their own.

Patterns, from least to most complex

  1. One primary provider and another for specific workloads. Each workload lives entirely in one cloud. It is the most common pattern and the healthiest.
  2. Connected hybrid. Your own infrastructure for what has to stay, cloud for the rest, with a private and secure connection between them.
  3. Portability by design. Containers, infrastructure as code and standard services so workloads can be moved if needed, accepting that you give up some provider-specific services.
  4. Active-active across providers. The same workload in two clouds at once. Only when the continuity requirement demands it, because it multiplies cost and complexity.

What has to be solved in every case

  • Unified identity: a single way of knowing who is who across every environment.
  • Network and connectivity: private, encrypted paths, with well-defined zones.
  • Common governance: the same tagging, security and cost rules in every cloud.
  • Cost in a single format, so it can be compared and added up (which is what FOCUS solves in the FinOps series).

How we apply it: Cloud Architecture.

If your cloud cost keeps growing and nobody knows why, or your architecture depends on whoever built it, let’s talk.

Let’s talk