NUI Galway – UL Alliance First Annual ENGINEERING AND - ARAN ...

1. Situation
Querying Live Linked Data
Jürgen Umbrich, Aidan Hogan, Axel Polleres, Stefan Decker
DERI, National University of Ireland, Galway
email: firstname.lastname@deri.org
A growing amount of Linked Data enables advanced
data integration and decision-making applications. Typical
systems operating on Linked Data collect (crawl) and preprocess
(index) large amounts of data, and evaluate queries
against a centralised repository. Given that crawling and
indexing are time-consuming operations, the data in the
centralised index may be out of date at query execution
time. An idealised query answering system for Linked
Data should return live answers in a reasonable amount of
time, even on corpora as large as the Web. In such a live
query system source selection - determining which sources
contribute answers to a query - is a crucial step to avoid
unnecessary fetch operations.
Most current approaches enabling query processing over
RDF data are based on materialised based approaches
(MAT). Centralised approaches provide excellent query
response times due to extensive preprocessing carried
out, but suffer from drawbacks like the freshness of the
aggregated data or data providers have to give up sole
sovereignty on their data. On the other end of the spectrum,
distributed query processing approaches (DQP) typically
assume processing power attainable at the sources themselves,
which could be leveraged in parallel for query
processing. Such distributed or federated approaches offer
up-to-date query results, however, they cannot give strict
guarantees on query performance and the response time is
very slow compared to MAT approaches.
Our current research direction is based on 1) the investigation
of a middle ground approach between the
two extremes of MAT and DQP and propose the use of
histogram-based data summaries [1] and 2) the studies of
the dynamics of Linked Data sources [2], [3]. The results
clearly show that a query approach which combines MAT
and DQP allows a novel and powerful way to guarantee
fast query response times and up-to-date answers. The
remainder of this abstract highlights the next steps to
enable novel ways to querying live Linked Data.
2. The System
Figure 1 depicts the proposed solution architecture
which consists of three main components: 1) the query
processor, 2) detecting dynamic patterns and 3) the source
selection components. The system can be plugged on top of
any RDF store which allows data access via the SPARQL
protocol. The query processor decides during the runtime
which parts of the query involve dynamic information and
122
live results
SPARQL
Query
Processor
non-dynamic
query patterns
dynamic
query patterns
MAT
detect dynamics
LinkedData
Web
source
selection
Any local or
remote index
providing data
access via
t h e S P A R Q L
protocol
Fig. 1. Combined Query of RDF store(s) and the Linked Data Web.
which are rather static. Clearly, the system will evaluate
static query parts with the MAT approach over the local
index and dynamic parts with an DQP approach directly
over the Linked Data Web.
3. Dynamic Patterns
Our current research heavily focuses in detecting dynamic
patterns in Linked Data sources which are then
incorporated into the query processor and the source selection
component (cf. “detect dynamics” in Figure 1). As
a direct consequence, the next research step is to integrate
the source selection component and the knowledge about
dynamic patterns into the query processor to decide at
runtime which parts of a query are evaluated over which
sources (either local or Web sources). One of the crucial
steps is to decide if we base our access decision on the
knowledge about dynamic patterns or static patterns and
how to represent such patterns.
4. Acknowledgement
The work presented in this paper has been funded
in part by Science Foundation Ireland under Grant No.
SFI/08/CE/I1380 (Lion-2).
References
[1] J. Umbrich, K. Hose, M. Karnstedt, A. Harth, and A. Polleres,
“Comparing data summaries for processing live queries over linked
data,” World Wide Web Journal, vol. accepted, 2011, to appear.
[2] J. Umbrich, M. Hausenblas, A. Hogan, A. Polleres, and S. Decker,
“Towards Dataset Dynamics: Change Frequency of Linked Open
Data Sources,” in Proceedings of the Linked Data on the Web
Workshop (LDOW2010) at WWW2010, 2010.
[3] J. Umbrich, M. Karnstedt, and S. Land, “Towards Understanding the
Changing Web: Mining the Dynamics of Linked-Data Sources and
Entities,” in Proceedings of the Knowledge Discovery, Data Mining,
and Machine Learning Workshop (KDML 2010) at LWA 2010, 2010.