1. Grid Computing

The current global trend in scientific research shows increasing emphasis on two aspects: (1) collaboration and use of distributed resources; and, (2) use of high-speed networks. Grid computing is an emerging technology that is gaining ground in collaborative and computational-intensive research activities because of its potential to be efficient in terms of cost and resource usage. The concept of Grid computing has three important and unique aspects. One, allows for distributed computing by bringing together geographically and organizationally dispersed computational resources to undertake a single problem using high-speed networks. Second, enables collaboration among researchers and institutions working on a common goal or project, making research efforts more efficient and integrated. Third, makes more cost-effective use of a given amount of computer resources since all devices connected to the network can be treated as virtual resources that can be shared.

In the Philippines, Grid computing is relatively in its initial phase. However, the concept of collaborative research using advanced network technologies has been advocated by ASTI, which operates the Philippine Research, Education, and Government Information Network or PREGINET, the country's National Research and Education Network (NREN). Envisioned to be a catalyst for research and education collaboration among government, academic, and research-based institutions, PREGINET now has over 80 local partners nationwide. Moreover, PREGINET is directly connected to regional and international RENs, including the Asia Pacific Advanced Network (APAN), and Trans-Eurasia Information Network 3 (TEIN 3), as well as participates in the Asian Internet Interconnection Initiatives (AI3), an open Internet testbed initiated by the WIDE Project of Japan.

With PREGINET  already established, the focus now of ASTI's continued efforts in advanced networking is on the development of applications that can run over the Grid. The PSiGrid Program, using a multi-institutional approach, targets to deploy applications in the life and physical sciences, including bioinformatics, meteorology, oceanography, and computational physics. Specific fields that can benefit from Grid computing technologies include, Earth Monitoring, Disaster Management, and Climate Change.

In pursuit of national development, the PSciGrid Program, which runs for 3 years, aims to establish a national Grid infrastructure as well as applications that will run over the grid for use in collaborative research activities.

2. High-Performance Reconfigurable Computing

High-Performance Reconfigurable Computing (HPRC) is a relatively new computing paradigm that combines the use of conventional microprocessors and programmable logic devices called Field-Programmable Gate Arrays (FPGAs) in compute-intensive applications.  FPGAs are commodity digital integrated circuits (ICs), the logic configuration of which can be programmed repeatedly depending on the target application.  Inside an FPGA is an array of programmable logic blocks and flexible interconnects that can be configured to create arbitrary digital circuits.  This is in complete contrast with traditional microprocessors or CPUs that contain hardwired digital electronic circuits that are fixed.

In a configurable computing system, the FPGAs serve as auxiliary processors to the microprocessors.  The main application software or program is executed one instruction at a time in the microprocessor while functions that usually require long execution time are implemented as a digital circuit in the FPGA.  Essentially, the FPGAs provide hardware-accelerated solutions to software-only implementation of functions.  Example of functions that are typically implemented as hardware are data-parallel overlapped computations that can be efficiently implemented as fine-grained architectures such as single-instruction, multiple-data (SIMD) engines, pipelines, or systolic arrays.  The use of FPGAs in high-performance computing is observed to improve execution performance by up to four orders of magnitude.    And since FPGAs are low-power devices, FPGAs also offer improvement in power efficiency by up to three orders of magnitude reduction compared to conventional microprocessors.

ASTI’s high-performance computing cluster currently employs FPGAs to accelerate some commonly used bioinformatics algorithms which include Smith-Waterman, ClustalW and HMMer.