A Cloud-Based, Geospatial Linked Data Management System - Transactions on Large-Scale-Data-and Knowledge-Centered Systems XX - page 86

to drop some queries when a particular and quite small limit of available main
memory is reached. Such scaling factors could be combined with the existing
ones in order to create more up-to-date scaling conditions which could involve
conjunctions or disjunctions of fine-grained conditions involving these factors
and their respective thresholds. It will also be investigated how Amazon could
support such more complex adaptation conditions as the current possibilities
offered include just the independent evaluation of fine-grained scaling conditions.
The following future work directions are envisioned. First, further enhancing
the reliability level of the system. Second, enabling full migration to different
cloud providers to avoid lock-in. Third, further enhancing the functionality of
LMS service by also enabling to cater even more advanced geospatial querying
scenarios as well as enabling the importing of different data formats apart from
RDF, XML and relational. Fourth, further enhancing the performance of even
the most demanding geospatial queries. Finally, an interesting research direction
to pursue would be to enable a bidirectional mapping between relational or
even XML data to RDF ones. Such a mapping would give the freedom to the
application or data provider to choose the way his/her data can be updated.
For instance, the provider might already have existing programs which update
a particular portion of his/her data in their original form but could also develop
new programs which, after moving towards LD, enable the updating of a different
data portion. As such, different updates on different data formats will be possible
with the capability to synchronize between them without also having to make
any modifications to existing data update programs.
Appendix A - Experiment SPARQL Queries
1st query
select ?earthquake ?id ?date ?asGML ?asWKT ?latitude ?longitude
?locstring ?doc ?docName ?depth ?magnitude where {
?earthquake a sci:S23_Earthquake;
crm:P7_took_place_at ?place;
crm:P1_is_identified_by ?id;
crm:P4_has_time-span ?date;
sci:O17_has_dimension ?dd;
crm:P70_is_documented_in ?doc.
?doc crm:P1_is_identified_by ?docName.
?place geo:hasGeometry ?point;
crm:P87_is_identified_by ?d;
crm:P87_is_identified_by ?long;
crm:P87_is_identified_by ?lat;
crm:P3_has_note ?locstring.
?d sci:O20_has_value ?depth.
?point geo:asGML ?asGML;
geo:asWKT ?asWKT.
filter(regex(?d,"/Depth/")).
filter(regex(?long,"/Longitude/")).
filter(regex(?lat,"/Latitude/")).
?dd crm:P90_has_value ?magnitude.