Sammendrag
Nowadays patients like to have the option of staying at home in their illness
period. This can be realized through observing and monitoring the patients
motion and temperature with an home care application. Sensors are placed in the
home and create a real-time stream which have to be analyzed.
Data Stream Management Systems (DSMSs) and Complex Event Processing
systems (CEP) support real-time analysis of data streams which is a continuous
ordered sequence of tuples. There are some differences in the design and
implementation of these systems. We analyze these differences and similarities
for two such systems; TelegraphCQ, a DSMS and Cayuga, a CEP system, with a
focus on how well they might suit the needs of home care application domain.
Some of the most important criteria for a system in a home care environment are
to have a system with a language that allows an application to formulate
appropriate queries. In consideration of this we conclude that compared to
Cayuga, TelegraphCQ is the most capable system for home care domain. The
reason for this is that the query language of Cayuga has too limited expressiblity.
And its operators are too difficult to use for expressing the queries that are typical
in this application domain.
TelegraphCQ in contrast does report the correct value and time interval with the
possibility to use built in functions from SQL and do not demand query with lot of
complexity.
The comparison is done in two parts; theoretical and practical. We focus on
topics like tuple definition, aggregation, consecutiveness, concurrency and
optimization. Through the first part we investigate the literature concerning the
two types of systems. The second part consist of testing these systems in the
home care application domain. This part allows us to test the topics discussed in
the theoretical part.