July 6, 2026 · Loïc Veyssière
Introducing Archipel Labs
Why a public laboratory for enterprise architecture, and how observing a simulated company beats explaining architecture from above.
Enterprise architecture has a habit of describing systems from above: through frameworks, meta-models, layers, taxonomies and standards. These tools are useful, but they explain architecture the way a map explains a country. At some point you have to walk the terrain.
Archipel Labs is a public laboratory built on that conviction. Instead of starting from a framework that tries to explain every possible architecture, it starts from a realistic simulated system: an architecture in motion, with applications, integrations, data flows, operational constraints and failures. Architectural patterns are then explained by observation, not only by abstract classification.
The vertical bet
There's a common caricature of the architect's role: someone who produces high-level designs and diagrams, at a comfortable distance from the code. In reality, many architectural decisions are deeply connected to implementation and integration details. Some choices simply cannot be understood from a high-level view alone. (The home page of this site makes the point in one click.)
That's why the lab follows a deliberately vertical approach: a core sample of architecture, drilled from strategic intent down to technical implementation, connecting the high-level picture with the real components, data flows and constraints that appear along the way.
The flagship: an e-commerce company
The first simulation follows an e-commerce company for a year. Today it is deliberately small: a real storefront selling Minecraft items, and customers simulated in real browsers. They browse, compare, abandon carts or buy, and the company produces the first signals of any business: traffic, orders, operational events.
It won't stay small. Suppliers, inventory, accounting and analytics will arrive one by one, each new capability adding pressure on the existing system, each integration creating new dependencies. The objective is a company in motion, never a finished demo. And e-commerce is only the first terrain: other kinds of businesses will follow (reservation, SaaS, manufacturing), each with its own architecture to grow and observe.
Why "archipelago"
An enterprise information system is an archipelago: many well-established packaged products (storefront, ERP, WMS, CRM, analytics), each one an island that works well on its own. The hard work of architecture is making these independent systems work together. It lives between the islands: orchestrating the flows, honoring the contracts, keeping the whole coherent. That is exactly what the lab makes observable.
Alongside the main simulation, the lab also runs self-contained experiments on one integration topic at a time, like the ServiceNow-driven Ansible automation blueprint: same philosophy, real products, real constraints.
I build Archipel Labs in the open, and this blog is its logbook: the learnings, the failures, the redesigns, with regular detours into theory. If walking the terrain appeals to you more than reading the map, follow the lab on GitHub or on LinkedIn, and come back for the field notes.
Archipel Labs · A public enterprise-architecture & data-systems laboratory.