Information Technology Reference
In-Depth Information
high speed of information. The management system also focuses on
online processing mechanisms for manipulating information such as
inserting and updating. In addition, it must provide features such as con-
sistency, integrity, security, and so on [17] like in a DBMS.
A major purpose of a data stream management system is to provide
users with an abstract view of the streaming data. That is, the system
hides certain details of how the data are selected and maintained.
Thereby, useful information data will not be lost, which offers an efi cient
retrieval, yet users see a simplii ed and easy-to-use view of the data. On
the other hand, a scientii c workl ow management system provides users
with an abstract view of the scientii c activities, processes, and their
external logic, hiding the details of how they are arranged and executed.
Considering that under the theme of the scientii c computation, most
activities and processes are concerned with dealing with raw data, the
processing of the scientii c workl ow is actually the processing of the data
l ow or data stream.
9.1.5.2
Data Stream View for Scientifi c Workfl ow
Inspired by the DBMS view for the scientii c workl ow in [6], the similarity
between data stream management and scientii c workl ow management
can also be found. In this subsection, the conceptual-level mapping
relationship between data stream management systems and scientii c
workl ow management systems will be presented.
For a given model of a certain scientii c workl ow, the execution of the
workl ow means the WFMS will manipulate and coordinate instances
from its creation to its cancellation. In other words, a model determines a
bunch of instances. This can be seen in
Figure 9.3
. As shown from the
vertical axis, the upper l ow is the workl ow model or workl ow schema.
Under the same workl ow model, during the execution time or runtime,
there are a series of instances processing different activities dei ned by the
model/schema simultaneously. But from the horizontal axis, it is obvious
that the input data l ows through the DSMS and the DSMS will match
the ongoing data with reference to the query pattern given at the top of
Figure 9.3. So it is clear that a scientii c workl ow schema can be treated as
a data stream query expression during the processing, which indicates
that the DSMS query processing technique can be applied to scientii c
workl ows.
We previously mentioned that in any DBMS, the access of the database
system depends on a unii ed access language that is used throughout the
dei nition, execution, and monitoring. This is also the case in the DBMS-
based system. In this sense, both DBMS and DSMS can improve scientii c
workl ow management.
Search WWH ::
Custom Search