[dna-dev] SOA/Enterprise project using DNA

[Stefano Maestri]

Randall Hauch wrote on 19/11/08 21:21:
>
> And, please feel free to suggest or discuss features that might be
> useful, too.  
The feature I'd like to see in a next future, not necessary for 0.4 of
course, is write connectors for graph api on DB and maybe on a
distributed storage (like hadoop). We will have a lot of data here, see
below.
> For example, do you think JCR versioning will be an important part of
> your approach, or will you implement your own approach to managing
> history by designing the hierarchy to handle it?
No it isn't for our impl. What we are going to do is to keep a snapshot
of service responses linked to a particular user request. We keep of
course accounting of user requests and we will use the primary key of
accounting system to keep versioning. (I hope I've been sufficient clear
in my explanation, if not just shout :) )
What we could need, not in the fisrt step of the project anyway, is
analyzer to extract different views of data already saved in JCR with a
particular structure. JCR is good because infos are already organized in
the way we will read them, but I'm not sure we will get perfect fit of
our needs at fist shot. So I immagine analyzers a way to re-organize JCR
in different manner (better).
> Also, do you have an idea about how much information the repository
> will store (perhaps broken down by % in uploaded files, and about how
> big those files will each be).
It is incremental increasing of course. Not a precise idea since we are
reviewing all the system to go to service. But we expect 1 terabyte in
18/24 months based on the experience of data we are storing on a DB. We
have a DB of pure data of 5 TB and we have an increasing of about 0.7
TB/year.  We are not moving all data on  JCR, but service data are more
expensive than pure datas....well it isn't a precise prevision...I'll be
more precise later in the project.

Bye
S.