George K. Thiruvathukal

Home networking

Although most of the ideas behind networking are relatively straightforward, several interesting technical and business decisions must be made as an ongoing part of the process. In this first installment of what I hope turns into a discussion about more advanced applications, the focus is on revealing the secrets in a typical home router/switch. (Here, a router has both switching and routing functions).

XML and computational science

The Extensible Markup Language (XML) is a specification for document interchange that the World Wide Web Consortium (W3C) developed in 1998. In many ways, XML is the lingua franca among programming language enthusiasts, and proponents argue that it could potentially solve the multitude of data management and analysis problems the entire computing industry currently faces. XML might make a real difference, especially in computing, engineering, and the mathematical sciences, in part because we can use it with different languages. The author presents some background and lightweight examples of XML usage, describes some XML component frameworks along with their purpose and applicability to computational science, and discusses some technical obstacles to overcome for the language to be taken seriously in computational science.

High-Performance Computing Applications on Novel Architectures

As chip designers struggle to increase the speed of conventional microprocessors, the scientific computing community is turning its attention to alternative architectures that utilize computational accelerators to improve the application performance. This special issue of CiSE examines several examples of computationally demanding applications implemented on novel high-performance computing architectures and points out the challenges and opportunities that these new architectures bring up.

Plone and content management

This paper looks at Plone, one of the best content management systems. Plone is distributed under a free open-source license: the cost of getting started is only limited to the time you have available to set up the software on a server. Plone is written in Python and uses the Zope application server infrastructure; it runs on most modern operating systems. Plone can be customized for the maintenance of content - entirely over the Web.

Java at middle age: enabling Java for computational science

Javas early history explains its lure to this day. Java is alive and well, running on just about every computing platform, from handhelds to high-end servers such as multiprocessors. Implementations of Java for Windows, Macintosh (including OS X), and Linux have all reached sufficient maturity and are in widespread use. Performance and resource usage remain a problem in most Java implementations, but the language is improving all the time. In this article, I examine the lure of Java for computational science, discuss the Java Grande effort to work with Sun, and identify areas for improvement. I focus on language and implementation issues that must be solved for Java to be taken seriously for computationally focused codes. As a motivational tool, I include a number of reflections on the C# language from Microsoft, which many claim is nothing more than Java with a new syntax. I excise a number of sections from the specification documentation to demonstrate that specific Java Grande Forum (JGF) recommendations are implemented in the current version of Microsofts C#. What is particularly interesting (but not necessarily verifiable) is that language appearing in the Microsoft C# specification proper often appears to paraphrase similar points raised in the JGF documents available at www.javagrande.org.

What I Did on My Summer Vacation

This article discusses the continuing importance of requirements, managing Web content, free public Wi-Fi and a personal transition to Internet telephony.

Scalable implementations of MPI atomicity for concurrent overlapping I/O

For concurrent I/O operations, atomicity defines the results in the overlapping file regions simultaneously read/written by requesting processes. Atomicity has been well studied at the file system level, such as POSIX standard. We investigate the problems arising from the implementation of MPI atomicity for concurrent overlapping write access and provide two programming solutions. Since the MPI definition of atomicity differs from the POSIX one, an implementation that simply relies on the POSIX file systems does not guarantee correct MPI semantics. To have a correct implementation of atomic I/O in MPI, we examine the efficiency of three approaches: I) file locking, 2) graph-coloring, and 3) process-rank ordering. Performance complexity for these methods are analyzed and their experimental results are presented for file systems including NFS, SGIs XFS, and IBMs GPFS

Natural XML for data binding, processing, and persistence

The article explains what you need to do to incorporate XML directly into your computational science application. The exploration involves the use of a standard parser to automatically build object trees entirely from application-specific classes. This discussion very much focuses on object-oriented programming languages such as Java and Python, but it can work for non-object-oriented languages as well. The ideas in the article provide a glimpse into the Natural XML research project.

Digging into Data

Computational projects such as Wordle can simplify our lives by extracting data and presenting it in a new, visually compelling way.

A Hike through a Post-EJB J2EE Web Application Architecture, Part III

By incorporating automated component, integration, and acceptance testing into the various tiers of a lightweight lava 2 Enterprise Edition (J2EE) Web application architecture, developers can shorten the development cycle and increase the quality of their work