Technical Requirements

Hardware

In order to run a repository locally (as opposed to buying a hosted option) you will require hardware. Repositories are typically run on servers housed in an institution's computer room which will be specially designed for the job (air conditioning, networking etc). Repositories can sit on dedicated servers, shared services (perhaps a shared web server) or as virtual machines on larger servers. Your local institution will have a policy for how they prefer to host services and servers. This policy make take into account storage and processor requirements, expected load on a server, or resiliency requirements (does it need to have a guaranteed 24x7 99.999% uptime?).

Most initiations, especially in the first few years when their repository is only sparsely populated) will find that a basic or moderately specified server will perform sufficiently for them. When considering your hardware requirements, consider:

How long you wish your server to last you for (3-5 years?)
How many items you think you are likely to deposit in that time period
What the average size of each item will be (to calculate your disk requirements)
What hardware specifications your chosen repository software recommends

Software functionality

Apart from the obvious computer equipment needed to set up a repository, e.g. server, storage space, network connection, it is the software that is installed and the web interface for users that determines what a repository looks like for most people. Functionality can vary depending on the software and what it was originally designed for (there are many communities who have developed repositories for their own specific purposes e.g. Libraries, E-science, E-learning, ). Typically, these functions may include:

Import
Export
Identify
Store
Get digital content
Describe the content and share the content / metadata

Metadata issues

[Information to follow.]

File formats

[Information to follow.]

Harvesting issues

[Information to follow.]

Interoperability & Integration issues

Interoperability and integration relates to the interworking of your repository with other systems using common standards and protocols. It is important that your repository interoperates with other systems. If it does not then you will not reap all of the rewards that using an interoperable system could.

Interoperability usually comes into play with the harvesting of your metadata for example using OAI_PMH, deposit via an interface such as SWORD, or discovery using a mechanism such as RSS feeds. Other systems you may run in your institution, or run externally by service providers will want to make use of these interfaces.

Integration between your repository and other institutional systems may be required. A common requirement is for your repository to delegate authentication to a local system, for example Shibboleth, LDAP or CAS. Some repository software solutions are able to perform this out-of-the-box, others will need extensive configuration or re-programming.