There could be many definitions for a dataspace. A more academic definition would be:
Dataspace is a joint effort that consists of loosely coupled, independent actors with equal rights that act as data producers, data stewards and data consumers in a business-to-business fashion.
A simpler definition is also described in our FAQ page: "Dataspaces are an abstraction of a group of companies and people sharing data for a common purpose".
This is not to be confused with data warehouses or data lakes, as dataspaces are not "housing data" instead they just represent a virtual space to add members and peers.
A dataspace may be useful when the needs of the organization to share data crosses the organizational walls. It doesn't mean that you cannot invite your colleagues, but if all members of the dataspace belong to the same company, then a centralized solution is not a political obstacle.
Within a dataspace other elements are enabled, such as service connections, commons and agreement templates.
Each user within a dataspace exists in the context of a corresponding organization (organizational unit).