Great that you raise this (while I work on my Semtech 2010 trip report).
Yes the pieces are indeed in place:
1. WebIDs
2. ACLs
3. HTTP based CRUD operations
4. Data Spaces (hosting one or more Named Graphs).
Microsoft via "Project Dallas" [1] are also gunning for the "Data as a
Service" market.
It works like this:
1. You curate structured data (in representations acceptable to Dallas
e.g. CSV or OData)
2. You get a Data Stall in the Dallas furnished mall (so till and other
fixtures and fittings are in place)
3. You load our Stall with Data (by registering and uploading your data
set(s) to Dallas)
4. Data Consumers purchase use of your Data via the till provided by Dallas.
The trouble with the model above is that it's still a form of lock-in
i.e., you have to go through a Microsoft mall.
With open cloud storage, we can make mini malls, and most important of
all, we can get very granular about the value unit (i.e. Data Item)
which I referred to as being like a "Cube of Sugar" during the Linked
Data bizmodel session at semtech 2010. We can sell sugar as on a per
cube, box, cartoon, truck, entire warehouse etc.. basis.
BTW - here's what constitutes the cube (six vital data quality factors):
1. Accuracy
2. Freshness (change sensitivity)
3. Navigability (what inference context rules give you)
4. Representation Formats (the more the merrier re. RDF model based
structured representations of data items)
5. Platform Independent CRUD mechanism
6. Provenance.
Each data item has an HTTP URI based Name which provides a degree of
branding and foundation for a lot of provenance related embellishments.
Each data item will live somewhere (e.g. in a Named partition in an HTTP
accessible Data Space).
Access to data items would be controlled by Data Space owners courtesy
of WebIDs and ACLs.
Data Space owner gets the HTTP logs (the gold dust re. computing use of
Data Items from ones Data Space).
Links:
1. http://pinpoint.microsoft.com/en-US/Dallas
--
Regards,
Kingsley Idehen
President & CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen
also worth noting that if you read between the lines (~ish) cisco is
pointing most of their business in this direction, products like cius
complete their cloud strategy.
also, and this is a *very* aside; the media industry is about to start
rolling out HDBase-T rather than HDMI for their cabling, which means HD
tv through Cat-6 cables, which means HD tv over the lan/wan - and
coupled with efforts like BBC's "project canvas" - means [1] in the near
future
[1] something related to linked (and web of) data I can't figure out yet
- but it's definitely something...
>> Each data item has an HTTP URI based Name which provides a degree of
>> branding and foundation for a lot of provenance related embellishments.
>> Each data item will live somewhere (e.g. in a Named partition in an HTTP
>> accessible Data Space).
>> Access to data items would be controlled by Data Space owners courtesy of
>> WebIDs and ACLs.
>> Data Space owner gets the HTTP logs (the gold dust re. computing use of
>> Data Items from ones Data Space).
>>
>> Links:
>>
>> 1. http://pinpoint.microsoft.com/en-US/Dallas
>>
>> --
>>
>> Regards,
>>
>> Kingsley Idehen President & CEO OpenLink Software Web:
>> http://www.openlinksw.com
>> Weblog: http://www.openlinksw.com/blog/~kidehen<http://www.openlinksw.com/blog/%7Ekidehen>
>> Twitter/Identi.ca: kidehen
>>
>>
>>
>>
>>
>