The UDC concept
[45] supports a layered architecture, separating the data from the application logic in the 3GPP system, so that user data is stored in a logically unique repository allowing access from core and service layer entities, named application front-ends. And such unique repository shall be possible to be shared among different PLMNs that have trusted relationships.
Network elements and functionalities should be designed to access profile data remotely and without storing them permanently locally, i.e. the front-ends shall work in a subscriber data-less configuration.
In some cases, services may depend on user data scattered over UDC and other network elements. UDC may support the ability to access necessary network elements to fetch user data on behalf of these services, while minimizing impacts on existing Network Elements in which the data is located.
Applications can subscribe to specific events which will likely occur on specific user data, and those should be notified when those events do appear. The events can be changes on existing user data, addition of user data, and so on.
Third-party applications and non-trusted network elements should only be able to access the user data after proper authentication and authorization taking into account security and privacy requirements, i.e. it should be possible to present different views on the data to the parties which require access, dependent on the authorization. UDC concept is backwards compatible with 3GPP systems, i.e. it does not have an impact on traffic mechanisms, reference points and protocols of existing network elements.
The UDC concept preserves user authentication and authorization of services across domains, ensuring secured users' access to network.
Due to the logical centralization of user data, it is necessary for UDC to support the provisioning of the user data, that is, user data manipulation like creation, deletion, reading, modification and other operations. Provisioning shall be possible via an external system, self-care or dynamically via applications offering e.g. user service configuration facilities.
Operations carried out in the framework of UDC shall support the ACID (Atomicity, Consistency, Isolation, and Durability) characteristics.
User Data Modelling refers to the different models that apply to user data convergence: Information Models and Data Models.
An Information Model denotes an abstract, formal representation of entity types, including their properties and relationships, the operations (e.g. read, write…) that can be performed on them, and related rules and constraints. In the information model, entities might have network topology relationship with each other.
In order to accommodate multiple applications and services, existing and new ones, a common baseline information model shall be developed and shall, at minimum, clearly distinguish a number of concepts as entity types:
-
Subscriber with relation to several users (e.g. a company and its employees),
-
A user attached to different subscriptions (e.g. for a private and a professional service usage)
-
A user using multiple devices (e.g. mobiles or fixed)
-
Grouping of users to certain categories
-
A particular user as member of a certain group
-
Service providers' services provided by network operators
-
Enterprise services provided by network operators
The baseline information model shall be future proof. It shall not be tied to any specific implementation of the data base or its interfaces. It shall provide flexibility (in its data structure and content), extensibility and multi-application approach.
By extensible, it shall be understood that new applications and/or new service profiles can be added by the operator, if necessary. The flexibility shall permit new data for existing applications to be introduced, or modified.
Data Models are practical implementations of the information model, e.g. Tree-like modelling. The common information model shall allow deriving one or more data models. A reference data model shall be standardized for the message exchange over Ud interface
[50], in order to enable multivendor interoperability.
Each application shall only interface the UDC for the data it is dealing with, and not be impacted by other data that UDC stores for other applications. It corresponds to the concept of a data view specific to a given application.
An application can allow access by other applications to data for which it is responsible. This can be done under certain constraint customized by operators.
Access to the UDC data shall be independent of the structure of the data models, i.e. the changes in the data models shall not affect the interface.