Abstracts of Magazine FUJITSU 2002-11 (VOL.53, NO.6)

Special Issue : Servers

• New Platform Concept: TRIOLE

To survive the increasingly fierce competition, many organizations have been actively expanding, merging or revamping their business systems. In spite of the resulting complexity, top management require their IT systems to respond more quickly and flexibly to unpredictable changes in the business environment. They also expect their systems to be more reliable, expandable, and offer lower total cost of ownership. Fujitsu is offering a new platform concept, TRIOLE, to meet these demands. This paper explains the three core technologies supporting TRIOLE: autonomy, virtualization, and integration. It also discusses how platform components such as servers, storage devices, networks, and middleware are effectively integrated and used under TRIOLE.

• UNIX Server: PRIMEPOWER Series

Fujitsu's PRIMEPOWER series of servers use a SPARC architecture and a Solaris operating environment and are scalable up to 128 CPUs. Now, Fujitsu has developed three new models: the PRIMEPOWER model 900 (with 16 CPUs), model 1500 (with 32 CPUs), and model 2500 (with 128 CPUs). These models have significantly improved performance, reliability, and flexibility because they use the highly efficient, highly reliable SPARC64 V CPU and other leading-edge technologies. They also feature a high-speed system clock and a substantially reduced memory latency due to centralized mounting of CPUs and memory modules. For the new models, Fujitsu drastically enhanced the component redundancy, the hot-swapping functions, and the resource-partitioning functions to improve the availability of the servers for mission-critical business systems. This paper describes the concept of the PRIMEPOWER series and the technologies used in the three new models to enhance their performance, reliability, availability, flexibility, and scalability.

• High-Performance Computer: PRIMEPOWER HPC

The PRIMEPOWER HPC is a new Fujitsu HPC server, which combines the vector parallel processing technology of the VPP Series with the symmetric multi-processor (SMP) technology of the PRIMEPOWER UNIX server. The PRIMEPOWER HPC enables an SMP node configuration consisting of up to 128 high-speed (5.2 GFLOPS) scalar CPUs to be built. As up to 128 processing nodes can be crossbar-connected via a high-speed optical interconnection unit, the maximum configuration of the parallel multi-node system attains a logical performance of 85.2 TFLOPS and 64 Tbytes of main memory. Features for high performance include high-speed CPUs, high-speed (133 gigabytes per second) snoop performance, and a barrier synchronization unit, while features for high reliability include redundancy of the node crossbar switch units and partitioning. This paper describes the parallel processing method and hardware features of the PRIMEPOWER HPC.

• SPARC64 V Processor for UNIX Servers

SPARC64 V is a newly developed processor for Fujitsu's PRIMEPOWER series of UNIX servers. It leverages high-performance and high-reliability technology that has been developed for mainframe systems. It conforms to the SPARC V9 architecture and supports Solaris Operating Environment, the de facto standard UNIX server OS. Tuned for high performance in the real-world enterprise computing environment, SPARC64 V implements a high throughput two-level cache; a 4-issue, 6-dispatch super-scalar pipeline; and an out of order execution of most types of instructions, including load/store requests. SPARC64 V achieves both a low power consumption and a high performance. It also satisfies mission critical application requirements with the latest mainframe level RAS features such as recovery of any 1-bit error in RAM and parity protection of most non-RAM data paths. This paper describes the micro-architecture overview of SPARC64 V.

• UNIX Server Software: PRIMECLUSTER

Enterprises have rapidly expanded their business activities, and the demand for IT-based business solutions has increased throughout the business world with the advent of the broadband age, the trend towards an around-the-clock service economy, and globalization. If these important systems become damaged or stopped for prolonged periods due to breakdowns, failures, errors, or malfunctions, business opportunities will be missed and serious social problems may even occur. Therefore, substantial measures must be taken to prevent and protect against such errors and malfunctions. PRIMECLUSTER is a high-availability cluster software product that protects businesses from such risks by managing system redundancy and maximizing system uptime. It also provides an absolutely unique business continuity when combined with Fujitsu's middleware products, PRIMEPOWER and ETERNUS GR series. This paper outlines the PRIIMECLUSTER development policy, a broad array of advanced technologies, and our efforts and approaches to providing customers with total business solutions.

• Intel-Architecture (IA) Server: PRIMERGY

"PRIMERGY"is the most cost-effective server produced by Fujitsu, with an Intel architecture that allows the Windows Server family and Linux to run on it. This sophisticated Intel-architecture (IA) server reliably meets a wide spectrum of market needs. In particular, the recent spread of broadband networks has fueled demand for Intel-architecture servers that can flexibly cope with rapid changes of demand. Fujitsu has therefore developed the blade server BX300, which completes the platform required for the broadband network age and meets the multi-server requirements of space saving, power saving, 24-hour operation, and increased operation and management efficiency. This paper describes the full lineup of PRIMERGY servers, including the BX300.

• Global Server: GS21

Fujitsu's GS21 600 Model Group are the high-end models of the GS Series, which are high-reliability, high-performance servers that form the backbone of enterprise information systems in the broadband Internet age. The GS21 models have a new gigahertz-range CPU that features cutting-edge 0.13 micron copper-wiring CMOS technology. A secondary cache was incorporated on the CPU chip to enhance memory performance, and a key cache was added to the memory control circuit to improve key storage access. A single cluster with 16 processors provides a high performance of 3,000 million instructions per second (MIPS). The reliability and availability of these models are assured with ECC protection of the inter-device data-bus interface and hot-system maintenance and expansion functions. This paper describes the features of the GS Series Global Servers and the hardware of the GS21 600 Model Group.

• Storage System: ETERNUS

In the broadband Internet age, storage systems will be separated from and independent of application servers and will be used as network storages shared among several servers on the network. With the network storage systems proposed by Fujitsu, data can be maintained autonomously, application servers can store and utilize the same data using both SAN and NAS, and necessary data can be accessed at any time from anywhere. The Fujitsu "ETERNUS" storage systems consist mainly of a network storage server, disk array, network disk array, and tape library, among which the network storage server is the heart of this integrated next-generation network storage solution. This paper introduces the development concept and actual product lineups of the ETERNUS storage systems.

• IPCOM Supporting Front-End Enterprise Systems

High-speed and wide-band networks, for example, the broadband Internet are becoming common and are evolving rapidly for realizing ALL-IP. Now, using the broadband Internet as an infrastructure, companies are increasing the scale and complexities of their enterprise systems to meet the needs for new businesses and strengthen intra-company and inter-company linkages. Under these circumstances, enterprise systems are having problems providing fault-tolerant continuous operation, handling the large changes in network load, and reducing the load due to complicated system operations. To address these problems, Fujitsu has released a new product called IPCOM that is based on the TRIOLE platform concept and has autonomous and virtual functions. IPCOM liberates data and processing by connecting the server and storage with the network without the user needing to consider the place and time, and achieves various integrations of enterprise systems. This paper describes the technology and functions of IPCOM using example applications.

• Forefront of Open-System Applications: PRIMEPOWER CENTER

Use of open technology systems, for example, UNIX, Windows, and Linux, has become essential to business promotion in the Internet era. On the other hand, system integration in a multi-vender environment is a large burden on integrators and users, because it requires prior checks and verifications. To ease this burden, Fujitsu has established the PRIMEPOWER CENTER. The PRIMEPOWER CENTER provides comprehensive support ranging from consultations about customers' systems to system demonstrations and verifications. The PRIMEPOWER CENTER also helps customers prepare software and hardware products and verify open technology systems in multi-vender environments through collaborations with Independent Software Vendor (ISV) and Independent Hardware Vendor (IHV) partners. This paper describes the features and activities of the PRIMEPOWER CENTER.

• User Case: Daiichi Pharmaceutical (SAP on PRIMEPOWER)

Daiichi Pharmaceutical Co., Ltd, a leading drug company in Japan, started production operation of its SAP R/3 system using a PRIMEPOWER server platform in April 2002. For this project, Fujitsu Ltd. supported the design, construction, and operation of the BASIS infrastructure of the system. Specifically, Fujitsu supplied high-performance, high-quality server, storage, and middleware products to ensure stable operation of Daiichi Pharmaceutical's new corporate core system and help the company meet an associated social commitment for timely distribution of medicines. In addition, as part of the cooperative work between infrastructure and applications, Fujitsu provided the high-quality support necessary for Daiichi Pharmaceutical to successfully perform the world's first implementation of the Production Planning Manufacture Execution System (PP-MES) module of SAP R/3. This paper introduces the project history of the SAP R/3 system's introduction in Daiichi Pharmaceutical. It also describes the selection/evaluation processes and the notable features of the system's platform.

• PRIMEPOWER System Supporting Fujitsu Tatebayashi System Center

The Internet is now an indispensable part of business management strategy. Fujitsu has therefore opened the Internet Data Center (IDC) to provide overall Internet support for businesses, ranging from planning to construction to business operation. Two main system centers have been established: the Tatebayashi Center in the east and the Akashi Center in the west. To build a multi-center system, Fujitsu also opened a Tokyo System Center in the autumn of 2001 that specializes in broadband. The management of these system centers maintains customer systems so they operate reliably and securely. The Tatebayashi system center has won the trust of customers thanks to its PRIMEPOWER UNIX server, which uses state-of-the-art technology. This paper introduces the IDC service and PRIMEPOWER-based service offered by the Tatebayashi system center.

• User Case on IA Server "PRIMERGY"

Because IA servers have recently been introduced into not only information processing systems but also mission-critical systems, the reliability, operability, and scalability of IA servers have become much more important. There is also growing demand for large-scale backbone systems based on IA servers which offer high efficiency and availability. To meet this demand, Fujitsu provides the IA server "PRIMERGY", which offers integration capabilities for advanced large-scale mission-critical systems. This paper describes a case study on a large-scale database system based on PRIMERGY installed in Meiji Life Insurance Company.

• 3.5-inch MO Disk: GIGAMO 2.3GB

Removable media for saving data reliably and exchanging data quickly is becoming increasingly important because of the growing size of personal files. Many disks have been commercialized as removable media for this purpose, but no medium is superior to the MO in terms of speed and reliability. In 1998, we developed a 1.3-Gbyte, 3.5-inch MO called GIGAMO through the use of magnetically induced super resolution. By combining this technology and land/groove recording, we achieved a 2.3-Gbyte capacity. We also developed a novel ID system. This paper describes the specifications of the 2.3-Gbyte MO and its performance in a practical drive. This paper also describes the results of a durability test and the reliability of the new GIGAMO.

• Original Documents Management System Using MO

The 3.5-inch MO has a long archival life of more than 50 years and is unaffected by sunlight, water, dust, and magnetic fields. Moreover, it can help prevent cyber terrorism because it can be easily removed from the MO drive. Because of these features, the 3.5-inch MO is becoming widely recognized as a reliable storage medium. Many of the documents used in companies and SOHOs have been recently digitized. According to Japanese law, certain documents, for example, accounting documents, must be archived for a long period without deterioration. Also, documents containing private information, for example, customer and hospital records, must be stored in a highly confidential manner. Fujitsu's MO Document Management System was developed for storing these important documents securely. The system protects data confidentiality by controlling access to the MOs, encrypting documents, and logging user actions so that unauthorized changes to the stored data can be detected. This paper describes the concept, configuration, method of use, and security technology of the system.