Any update on the performance benchmarks mentioned below? I
know Kowari has a new release expected soon (Dec?), and sesame
just released an alpha of sesame2. Thought an update of this
report, would be very informative and timely:
http://simile.mit.edu/reports/stores/
I have 2 big memory machines (64bit aix w/ 64GB RAM, and a
64bit DebianSarge dual core opteron w/ 16GB RAM - using the 64
bit IBM 1.4.2 on aix, and the 64 bit Sun 1.5 on sarge64)- I
have been interested in experimenting with the newer 3stores
for large in memory storage of bioinformatics data. I'm
currently experimenting with sesame and the uniprot data.
You also mentioned a future effort to address scalability
concerns - and the dedication of an engineer to this topic.
I'm guessing we will have to eventually start looking at how
to distribute graphs across a cluster - thoughts? Google's
work on 'BigTables' seemed very interesting. Happy to chat
about this topic some more some time.
If the work mentioned below is in process - let me know if I
can help with some of the benchmarking using the hardware I
mentioned.
Incidentally, my interest in the Simile project is currently
just SemanticBank/Longwell.
On another topic: I noticed an ajax branch in the longwell svn
repository. What are the plans for this? Have you looked at
some of the various js frameworks to support this development
(e.g. scriptaculous, dojo, zimbra's ajaxtk, etc.)?
best,
ian
Stefano Mazzocchi said the following on 9/26/2005 11:39 AM:
> Arjohn Kampman wrote:
>> Stefano Mazzocchi wrote:
>>> Richard Cyganiak wrote:
>>>
>>>> Dear list,
>>>>
>>>> I'm under the impression that Simile projects use or have used both
>>>> the Sesame and Jena frameworks. I'm in a discussion about the
>>>> relative benefits of both frameworks. I'm curious about your
>>>> experiences. In the context of Longwell and Piggybank and Semantic
>>>> Bank, what are the strengths and weaknesses of both APIs? Are your
>>>> experiences documented anywhere?
>>>
>>>
>>> The topic is complex and the feelings to hurt are many :-) so my
>>> diplomatic side would say "they both do their job well".
>>
>> We can handle it ;-)
>>
>> Seriously, one of the faster to ways to progress is to get criticized
>> for the things that are "not so good". All feedback is welcome, either
>> on- or off-list.
>
> Vineet and I will start working on a "triple store shootout" real soon
> now (as soon as Piggy Bank 2.1.0 is released)
>
> The results will hopefully be easily reproducible by others (on other
> platforms/machine combinations) and easily comparable (I'm not a fan of
> absolute performance evaluations, relative ones are much more useful,
> IMO) and we'll make everything publicly available, hoping to help point
> out weaknesses and strengths of the various triple stores.
>
> (and to show us and our users where the scalability walls are and how to
> route around them, especially with real life data).
>
Received on Tue Nov 08 2005 - 11:06:09 EST