Cyberinfrastructure in New York State - Grid Computing: Infrastructure, Service, and Applications - page 43

2.3.5 Benchmarking SOAP and XML

Web services-based specii cations have emerged as the underlying architec-
ture for core grid services and standards, such as WSRF. XML and SOAP, and
are inextricably intertwined with Web services-based specii cations, and as a
result the design and implementation of SOAP and XML processing tools
play a signii cant role in grid applications. These applications use XML in a
wide variety of ways, including workl ow specii cations, WS-Security-based
documents, service descriptions in WSDL, and on-the-wire format in SOAP-
based communication. The application characteristics also vary widely in
the use of XML messages in their performance, memory, size, and processing
requirements. Numerous XML processing tools exist today, each of which is
optimized for specii c features. To make the right decisions, grid application
and middleware developers must, therefore, understand the complex depen-
dencies between XML features and the applications. We have developed a
standard benchmark suite for quantifying, comparing, and contrasting the
performance of XML and SOAP processors under a wide range of represen-
tative use cases (Head et al., 2005, 2006). The benchmarks are dei ned by a set
of XML schemas and conforming documents. To demonstrate the utility of
the benchmarks and to provide a snapshot of the XML and SOAP implemen-
tation landscape, we have conducted a performance study of many different
XML implementations on the benchmarks.

2.3.6 Component-Based Grid Computing

An important challenge in building and deploying high-performance
scientii c applications is providing a software development model that
abstracts the complexity of the runtime environment and simplii es the
task of scientists, allowing them to focus on the details of their particular
application. A consortium of university and national laboratory research-
ers launched the CCA Forum to develop a common component architec-
ture (CCA) specii cation for large-scale scientii c computation. The CCA
specii cation dei nes the roles and functionality of entities necessary for
high-performance component-based application development. The spe-
cii cation is designed from the perspective of the required behavior of
software components. However, the design and implementation of the
framework, choice of communication protocol, and component contain-
ment mechanisms have not been formally specii ed. This has facilitated
different research groups to design, develop, and evaluate the use of the
same CCA specii cation to support a wide variety of applications. We have
designed and implemented a C++-based CCA framework, XCAT-C++,
which can efi ciently marshal and unmarshal large datasets, as well as
provide the necessary modules and hooks in the framework to meet the
requirements of distributed scientii c grid applications (Erdil et al., 2005;
Govindaraju et al., 2005, 2007).