Operating Repositories and Services in the NRP

This documentation covers use cases of running repositories and related services in the National Repository Platform (NRP).

In general, the reader should be familiar with current conditions for creating repositories in the NRP that describe the process from organisational view.

Note that you are not expected to handle establishing a repository yourself, you should always contact repository system specialists first.

In this document, we describe a more technical level, we discuss procedures and models of dividing responsibilities in operation of services in the NRP. The basic division is whether you intend to use core repository systems (i.e. CESNET-Invenio, CLARIN-DSpace, ASEP/ARL) or a specific implementation of a repository or of a related service (we told you that you need to read the conditions linked above first!). In addition to that, there are subtleties depending on the level of customisation your user group requests for your repository even in case of core repository systems.

Generic NRP Architecture

To discuss its usage at this level, it is necessary to understand the basic architecture of the NRP.

The lowest layer of the infrastructure consists of hardware clusters. The clusters are operated by the NRP staff and support two major enabling technologies for the whole enterprise: S3 storage and Kubernetes clusters.

In a nutshell, repositories and services run in Kubernetes and store data into a “virtually reliable” S3. If you are interested in principles how this level of reliability is achieved, check the detailed description of NRP storage and containers.

Involved Personnel, Roles, and Responsibilities

The storage and containers of the NRP are operated by the infrastructure personnel, directly reachable through L3 and storage support channels.

The repository itself is managed by the repository administrator (you have read the conditions linked on the top, haven’t you? And note that the repo admin is a role and responsibility, not a single person!). Who exactly operates the technical part of the repository depends on many details. We provide a generic overview. Keep in mind that the details of the setup always need to be carefully negotiated for individual cases.

Running Repositories Based on Core Implementations

If you intend to run an instance of a core repository implementation supported by the NRP, the infrastructure supports various levels of customisation of the repositories that directly translate to division of responsibilities between the NRP infrastructure and the repository administrator. Continue to the documentation on running standard repository implementations in the NRP for further details.

Running Specific Applications and/or Other Repository Implementations

For running services and repository implementations that are not directly supported by the NRP, the infrastructure offers the S3-based storage and environment to run containers. The repository (or service) administrator must ensure the technical operation of the repository itself (including sufficient personnel to prepare the deployment and to guarantee a reasonable quality of service throughout the whole life cycle).

To establish a service in this mode, kindly refer to documentation on running services in the NRP.