Re: TPF and DBMSes (was Re: Hydra and Shapes)

On 11/24/14 3:30 AM, Ruben Verborgh wrote:
>>> Hence, we are able to host them through the HDT compressed triple format,
>>> which gives excellent performance for those cases,
>>> far better than what I've seen any DBMS do.
>> How can that be so?
> Because HDT is much simpler than a DBMS;
> it can only do read-only operations. Everything is static.

Simple question: are you inferring that SPARQL query solutions against 
your HDT based static datatet are faster than Virtuoso against the same 
static dataset? I really need to know since this is the first time I am 
hearing you make such a claim.

>
> We have measured that HDT performs the combination of
> “looking up data corresponding to a triple pattern”
> plus “give an estimate count of the number of matching triples”
> faster than Virtuoso or any other DBMS we tested.
> For numbers, see http://linkeddatafragments.org/publications/ldow2014.pdf#page=7.

Why a PDF? You should have a collection of SPARQL Protocol URLs to 
backup up your claim. Then we have something that we can respond to, 
with ease :)


>
>> As far as I know, you haven't made claims about trumping a SPARQL DBMS in the performance stakes.
> I indeed did not.
>
> If you look at my original reply, out of which the above snippet was taken,
> you will see it was in the context of triple pattern fragments (the “them” above).
> I said that HDT is faster than all DBMSes I have tested
> for serving triple pattern fragments—but an HDT file is static.

BTW -- can you share with me, via a URL, an example of a multiple 
statement triple pattern fragment?

>
> Best,
>
> Ruben
>


-- 
Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this