Re: Restpark - Minimal RESTful API for querying RDF triples

On 4/17/13 2:26 PM, Jürgen Jakobitsch wrote:
> "My yardstick simply involves reliably servicing valid SPARQL queries
> with a valid response over valid large inputs. Can you do useful things
> with SPARQL at scale? Yes. Can you support a fully compliant SPARQL
> engine at scale? No. (Again, "I give up" or "here's a partial response"
> are not valid SPARQL responses.)"

To clarify my comments, we have an "anytime query" feature that 
basically brings the issue of scale and computing resources close 
together. This feature applies to SQL and SPARQL and it uses a 
quiz-style metaphor:

1. Query is posed
2. Query processor attempts to produce a solution within a default query 
timeout, that's extendible via retries.

If the query processor cannot produce a complete solution in the 
allotted time it produces a partial solution and an indicator that said 
solution is partial. Then there's an option to retry i.e., get more 
time, until the solution is complete. Thus, you ultimately you get a 
solution, but production time is dependent on computing resources and 
data set size etc..

This is all about scalability ultimately being out of the hands of 
SPARQL and placed back into the hands of computing resources (which 
includes: the SPARQL processor, host operating system memory, and 
hardware components such as CPU, disk, interconnects etc..).

> would you also say that a ferrari is a slow car if you see it cruising
> in a speed limit zone?

Exactly!

>
> just started to read this paper "The Complexity of Evaluating Path Expressions in SPARQL" [1]
> i'll be back in seven and a half million years :)
>
> wkr turnguard
>
>
> [1] http://www.theoinf.uni-bayreuth.de/download/pods12submission.pdf
>
>
>
Links:

1. http://bit.ly/11w8nTW -- old post about Virtuoso Anytime Query (this 
is a V6 feature which carries over to V7)
2. http://bit.ly/YvC1py -- tips and tricks note
3. http://bit.ly/11h7bBk -- old technical white paper about infinite scale.

-- 

Regards,

Kingsley Idehen 
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Received on Wednesday, 17 April 2013 18:58:40 UTC