The purpose of this service is to provide a mechanism for temporarily but robustly storing files such that a file can be passed through an ALF application by reference and retrieved for use some time after its creation. While the function of this service could be met using a networked file share that is commonly accessible, this service encapsulates the implementation providing amongst other things, the possibility of an intermediary to move the file between disjoint contexts.
There are two models for this service. They can co-exist.
1. As a specification that is implemented by tools that wish to share documents
In this model there would be many stores and files would be stored locally to the tools that produce them. These tools would hand out file references (eg urls) and assure some reliable lifespan for files thus exposed.
2. As a centralized common service that tools upload files to.
In this model a centralized common service is provided. Tools that create files can be made aware of the presence of the service and upload any file they wish to share. The common serivce would would hand out file references (eg urls) and assure some reliable lifespan for files thus exposed.
The "dowload" aspect of these two models is identical in function. The operational difference is one of multiple sources of varying capability versus a single source of predictable capability.