Repositories, replication and correction

The objective in this section is to set out requirements, possible approaches, and applications.

Need: to be able to aggregate, in a central location, all the information about myself, e.g. the trail left in various information systems of organizations I've been dealing with, from kindergarten to nursing home (and, eventually, I want to reduce the size of this trail to its minimum!) - this requires standardized data format as well as protocols (e.g. SAML) ensuring that I control who has access to what piece of information.

Need: an e-portfolio resilient enough, so when a primary source of information disappears or when technical standards evolve, I am still able to access it, e.g. access a record from a startup that went bust - this requires mechanisms such as caching (the type of service offered by Google) and replication.

Need: content identification: reference scheme allocated centrally, holding references to internal identifiers in repositories. Repository identifiers are never revealed to outside world, meaning that externally, EPMS appears to be the database.

Strategies:

• EPMS to manage as transparently as possible the connection with different databases/repositories, so that the everyday user just sees a virtual database/repository.
• EPMS holds table of contents with tables connecting each item with the chosen storage locations for it. Repository identifiers are never shown to outsiders.
• Sensible option to have one storage location local to the EPMS.
• Facilitate dedicated backup.
• Enable use of storage options that have physical update: may need to send e-mail or letter to human.
• Authority and replication policy: if more than one storage option chosen, nominate authority order of stores. Distribute update requests in turn. If update fails, or until confirmation received, the store sinks down the list of authority. Maintain ideal order, and put in place a continuous process to check that real-time preference order is harmonised with ideal order.
• All access by third parties is through the EPMS, not directly to any of the repositories. This is essential to give proper control over access. More detail in permissions and access control.

Possible applications:

• Personal information and file backup utility.
• Universal contact updater, including write-only moderated update access for general suppliers.

Questions:

• Are the suggested strategies adequate to meet the needs outlined?
• Are there likely to be any major obstacles preventing the adoption of any of these strategies?