Implementing DWDM Lambda-Grids

Abstract

Unprecedented demand for ultrafast and dependable access to computing Grids contributes to the accelerating use of dense wavelength division multiplexing (DWDM) technology as a Lambda-Grid enabler. In the Lambda-Grid space, the DWDM infrastructure provisions dynamic lambdas or wavelengths of light ondemand to support terabyte and petabyte transmission rates; seamless access to large-scale aggregations of feature-rich resources; and extendible Grid and inter- Grid services with predictable performance guarantees (Boutaba, Golab, Iraqi, Li, & St. Arnaud, 2003). DWDM Lambda-Grids consist of shared network components that include interconnected federations of other Grids, dense collections of computational simulations, massive datasets, specialized scientific instruments, metadata repositories, large-scale storage systems, digital libraries, and clusters of supercomputers (Naiksatam, Figueira, Chiappari, & Bhatnagar, 2005). As a consequence of the convergence of remarkable advances in DWDM technology and high-performance computing, Lambda-Grids support complex problem resolution in fields that include seismology, neuroscience, bioinformatics, chemistry, and nuclear physics. This chapter begins with a discussion of Grid development and DWDM technical fundamentals. In the sections that follow, the role of the virtual organization (VO) in establishing and supporting DWDM Lambda-Grid initiatives; capabilities of the Globus Toolkit (GT) in facilitating Lambda-Grid construction; distinguishing characteristics of Lambda-Grid operations, architectures, and protocols; and major Web services (WS) specifications in the Lambda-Grid space are examined. Descriptions of DWDM Lambda- Grid initiatives and security challenges associated with DWDM Lambda-Grid implementations are presented. Finally, trends in DWDM Lambda-Grid research are introduced.