A Data Stream View of Scientific Workflow - Grid Computing: Infrastructure, Service, and Applications - page 227

It should be mentioned here that the graphic patterns of workl ows are
not limited to the above structures. According to Workl ow Patterns by van
der Aalst [14], there are in total 20 patterns of workl ow schema such as
sequence, parallel split, and so on.
From the dei nitions of those patterns, we can deduce that a workl ow
schema can be simulated by the DFA [18,19]. It indicates implicitly that if
there is any technique highly related to DFA and automata theory it may
be used for processing the issues over the workl ow. On the other hand,
the technique should have a good capability for data processing because
scientii c workl ow makes its reputation for the high data intensity.
Consequently, combining the data processing ability and automata theory
result from data streams can provide an alternative view for scientii c
workl ow from a new angle.

9.2.2 Schema in DBMS and DSMS

Generally speaking, the schema describes the structure of a database
system in a formal language supported by the DBMS. In a relational data-
base, the schema dei nes tables, i elds in each table, and the relationships
between i elds and tables. Schemas themselves are generally stored in a
data dictionary by DBMS. Although a schema is dei ned in text database
language, the term is often used to refer to a graphical depiction of the
database structure.
However, whatever the system is, DSMS or DBMS, under the relational
data structure, all the data are arranged using tables and i elds, which
constrains the patterns of the relational schema. It is obvious that the
relational schema is totally different from the graphic/automata-based
workl ow schema. The data processing technique cannot be easily changed
and deployed for the scientii c workl ow processing.

9.2.3 XML Schema and Semistructured Data

Nowadays, XML-format Web data are getting popular and XML data
sources are of marvelous diversity [16,20,21]. Available XML data ranges
from small Web pages to ever-growing applications, such as biological
data, astronomical data, commercial data, and even to rapidly changing
and possibly unbounded streams that are often encountered in Web data
integration and online publish-subscribe systems. In May 2001, the W3C
released the XML schema standard [10]. This standard had a very long
gestation and this is not surprising, as the aim was to create a single mod-
eling language that would please all interested parties.
In terms of features, XML schema models are backward compatible
with DTDs. This is very important for the practical reason that it eases
the transition from DTD modeling to XML schema modeling. It is always
possible to convert a DTD into an XML schema model automatically. At
the same time, the XML schema can offer the complement to the XML