This is a brief guide to using LOCKSS in an environment where content is stored directly into the system rather than having it collect content from websites. We expect many PLNs will adopt this usage pattern.

Traditionally Archival Units have been defined by the characteristics of the content they collect: one or more starting URLs plus crawl rules which determine which URLs reachable from the starting points should be included. Each box in a network then collects (or should collect) the same set of URLs for the AU, which allows the polling protocol to arrive at a consensus for what belongs in the AU, and to detect and repair anomalous content.

Named Archival Units provide an alternative means of defining AUs, There are no start URLs and no crawl rules; instead content is stored directly into the LOCKSS repository by PLN infrastructure. But the essential characteristic of traditional AUs must be maintained - the copy of an AU on each box must have the same identity and the same content, in order for polling to work. And now it's the responsibility of the PLN admins and infrastructure to ensure this consistency. That means the using the same handle for the AU on each box and storing the same set of files in the AU.

Another difference is that, with crawled AUs, the system doesn't start polling the AU until it has successfully completed at least one crawl. That gating mechanism doesn't work with Named AUs; once the AU is created (configured) it may poll at any time, hence the order of operations below.

To create a Named AU follow these steps

1. Choose a handle for the AU and obtain its AUID using POST /auids on any box in the PLN.
2. Store all the files, or the first set of files, in the repository of each box using POST /artifacts.
3. Configure the AU on each box using PUT /aus/{auid}, passing it the AUID and an auConfiguration comprising the AUID and a map with one entry whose key is "handle" and whose value is the chosen handle.