[Logo Microbiome Studio]
STEVEN BLANCHARD
Function
_
steven.blanchard@microbiome.studio_
MICROBIOME STUDIO BY ABOLIS BIOTECHNOLOGIES
Station F - 5 parvis Alan Turing
75013 Paris, France
_ https://microbiomestudio.com/_
_Ce message ainsi que les pièces jointes sont établis à
l'intention exclusive de ses destinataires et sont confidentiels. Si
vous recevez ce message par erreur, merci de le détruire et d'en
avertir immédiatement l'expéditeur
This message and any attachments may contain privileged and
confidential information. It is intended solely for the addressees and
is confidential. If you receive this message in error, please delete
it and immediately notify the sender._
Le 2024-07-11T14:52:16.000+02:00, Andy Seaborne <a...@apache.org> a
écrit :
> On 11/07/2024 11:47, Steven Blanchard wrote:
>> Dear,
>>
>> We want to improve our fuseki instance to reduce the request time
>> when
>>
>> we launch multiple request at the same time on the same dataset.
>>
>> We launch our request by the fuseki REST API.
>>
>> We have the impression that the requests are processed one after
>> the
>>
>> other.
>
> Fuseki is multiple reader and single writer
>
> There can be as many readers as you like (limited by hardware) and
> they
>
> all execute in parallel. One thread each.
Ok, perfect. This is the default configuration: the fuseki server
determines the number of threads available.
> In addition, there can be a single writer and further writers are queued.
>
>> We tried to create multiple fuseki instance that use some commun
>>
>> dataset but we have a read acess error when we start the second
>>
>> instance because of the first instance create a lock file in the
>>
>> dataset.
>>
>> It's possible to do this or it's a fuseki limitation ?
>
> If the first part is sufficient,
>
> see the non-Apache project RDF Delta
> https://afs.github.io/rdf-delta/
>
>> They have some other tips, advices or possibility to improve the
>>
>> multiprocessing capacity of a fuseki server? to reduce time of
>>
>> request?
>
> What do the requests look like?
>
> The shape of requests greatly affects the answers here.
For the most part, these are queries to a large database where we want
to extract informations linked to a specific URI.
In this case, we put the triplet with the subject on which we want to
filtering (with VALUES) first. These queries return few results.
We have also some request where we want to filter on a string
predicate and we extract all the URI linked to this string. These
queries can return a lot of results (We have seen that query time is
correlated with the number of results).
> Andy
>
>> Our final goal are to upgrade our "developpement" fuseki server
>> with
>>
>> default paramaters to a production server with an high
>> scalability and
>>
>> optimised response times.
>>
>> Thanks for our help,
>>
>> Steven
>>
>> [Logo Microbiome Studio]
>>
>> STEVEN BLANCHARD
>>
>> Bioinformatics_
>>
>> steven.blanchard@microbiome.studio_
>>
>> MICROBIOME STUDIO BY ABOLIS BIOTECHNOLOGIES
>>
>> Station F - 5 parvis Alan Turing
>>
>> 75013 Paris, France
>>
>> _ https://microbiomestudio.com/_
>>
>> _Ce message ainsi que les pièces jointes sont établis à
>>
>> l'intention exclusive de ses destinataires et sont confidentiels.
>> Si
>>
>> vous recevez ce message par erreur, merci de le détruire et d'en
>>
>> avertir immédiatement l'expéditeur
>>
>> This message and any attachments may contain privileged and
>>
>> confidential information. It is intended solely for the
>> addressees and
>>
>> is confidential. If you receive this message in error, please
>> delete
>>
>> it and immediately notify the sender._
