[Sois-tcoa] File Transfer Service

[Scott Burleigh]

Stuart Fowell wrote:
> Dear all,
> 
> One of the actions from the Athens meeting was to determine the goal of
> the File Transfer Service by 1st May. So here's my kick-off of the
> discussion:
> 
> From my perspective, I see little scope at the moment for transfer of
> files between file stores on-board a spacecraft. How many different file
> stores are there?

I agree.

> However, there is the need for standardised distributed access by
> applications to some form of persistent storage. This is more than just
> a file store, as there are other mechanisms, e.g. FIFO/LIFO structures,
> linked lists (including packet stores), File Management System (FMS) for
> hierarchical and flat file stores, tape recorder and private partitions
> (user-defined format).

I am inclined to agree here too, but now we're talking about a much broader 
problem.  A standard service interface providing onboard shared access to a 
(notional) global filestore is something I think we could put together 
fairly quickly.  Any sort of general interface to all these different 
flavors of persistent storage would be a lot larger and more complex.  Are 
we sure there's an identified need for it?  Does it fall within the current 
TCOAS charter?

> FYI: There are two parallel studies, funded by ESA, looking into this.
> Their "objectives of the project are to define a Data Management
> Software architecture suitable to support all the payload and platform
> requirements in terms of data storage access for any type of mission,
> make the architecture independent from the mass memory technology and
> compatible with the COTS software used. In order to be able to cope with
> a wide range of project requirements, the system must be configurable
> and sizeable in term of capacity, data-rate organisation, interfaces and
> Mass Memory technology used. It also have to be reconfigurable in-flight
> in order to fulfil mission requirements in terms of availability,
> maintainability. Finally, optimisation of its power consumption is
> mandatory for future interplanetary missions as well as small satellite
> applications."
> So I propose that we should set a goal of providing a Persistent Store
> Access Service. This can be used as a building block for more
> format-specific services such as a File Access and Management Service,
> for which the file store requirements in CFDP provide a starting point.

If we do decide to move in this direction, there's pretty mature persistent 
storage system ("Simple Data Recorder", or SDR) built into JPL's 
implementation of CFDP that might offer a starting point.  I think it 
actually would meet most or all of the objectives of these ESA studies, 
it's been very thoroughly tested over the past 9 years, and it's flying 
today on Deep Impact.

Scott