Discovering Critical Factors Affecting RDF Stores Success
201
Table 1 List of evaluated criteria and corresponding case studies from which they have been elicited
ID
Criteria
Case studies
Is funct.
C1
Handling data in
motion
D, P
N
C2
Security of the graph All
N
C3
Versioning of the
graph
A
Y
C4
Handling multimodal
data in a knowledge
repository
A, P
Y
C5
Zero impedance
mismatch
D
Y
C6
(Distributed)
reasoning
P
Y
C7
Federation and
replication to support
data distribution
D, P
N
C8
Spatio temporal
database
All
N
D stands for D4All, A for Apps4ME, P for PEGASO
3.3 Analysis of the Critical Success Factors
3.3.1 Handling Data in Motion
In IoT based scenarios it is continuously growing the amount of live data in motion
(also known as streaming data), i.e. data continuously generated by different sources
(e.g. sensors, etc.) to be then lively transmitted, collected and analysed on a remote
point. These data in motion are typically produced at high frequencies, e.g. once
per millisecond or even higher in specific contexts. The processing of these data
provides both short term, closed-loop and live decision making capabilities, and
scalable long term off-line analysis capabilities. The advantage of using these data
is that the analysis is done on fresh data, which is temporal close to the event (e.g.,
cyber-security attack) or the condition (e.g., machinery in degraded status) that must
be detected. The disadvantage is that specific computing capabilities are needed to
process the data without delay [37]. In addition, the data variety (i.e., heterogeneity)
and their velocity (i.e., frequency of change over time) drive respectively to the
problem of the interpretation of these data and the management of the digital tools
performance. In particular, regarding the first problem, the lack of integration among
the data produced by various sources typically separate crucial streams of data and
increase the issue of too much data but not enough knowledge [32]. In order to
contribute to overcome this gap, it is relevant to understand if SWT (and in particular
RDF stores) are ready and mature to support data in motion.
201
Table 1 List of evaluated criteria and corresponding case studies from which they have been elicited
ID
Criteria
Case studies
Is funct.
C1
Handling data in
motion
D, P
N
C2
Security of the graph All
N
C3
Versioning of the
graph
A
Y
C4
Handling multimodal
data in a knowledge
repository
A, P
Y
C5
Zero impedance
mismatch
D
Y
C6
(Distributed)
reasoning
P
Y
C7
Federation and
replication to support
data distribution
D, P
N
C8
Spatio temporal
database
All
N
D stands for D4All, A for Apps4ME, P for PEGASO
3.3 Analysis of the Critical Success Factors
3.3.1 Handling Data in Motion
In IoT based scenarios it is continuously growing the amount of live data in motion
(also known as streaming data), i.e. data continuously generated by different sources
(e.g. sensors, etc.) to be then lively transmitted, collected and analysed on a remote
point. These data in motion are typically produced at high frequencies, e.g. once
per millisecond or even higher in specific contexts. The processing of these data
provides both short term, closed-loop and live decision making capabilities, and
scalable long term off-line analysis capabilities. The advantage of using these data
is that the analysis is done on fresh data, which is temporal close to the event (e.g.,
cyber-security attack) or the condition (e.g., machinery in degraded status) that must
be detected. The disadvantage is that specific computing capabilities are needed to
process the data without delay [37]. In addition, the data variety (i.e., heterogeneity)
and their velocity (i.e., frequency of change over time) drive respectively to the
problem of the interpretation of these data and the management of the digital tools
performance. In particular, regarding the first problem, the lack of integration among
the data produced by various sources typically separate crucial streams of data and
increase the issue of too much data but not enough knowledge [32]. In order to
contribute to overcome this gap, it is relevant to understand if SWT (and in particular
RDF stores) are ready and mature to support data in motion.