Interpretation of NEC's X86 fault-tolerant server technology

In the 1980s, the first generation of fault-tolerant technology began to enter the commercial field. American Stratus (Fault-tolerant) uses the Motorola M68000 processor in Stratus' unique hardware-level fault-tolerant technology and VOS proprietary operating system environment.

In 1993, the Intel I860 processor was successfully applied in Stratus' hardware-level fault-tolerant architecture. In the software environment, it also satisfies the industry's open-end Unix operating system FTX, namely AT&T UNIX SVR4. .

In 1996, fault-tolerant technology was supported by HP, and the Stratus Continuum series was introduced to combine Stratus fault-tolerant structures with HP PA-RISC symmetric multi-processing technology.

Since the beginning of the 21st century, the demand for servers, especially low-end and mid-range IA servers in manufacturing, SMEs, energy, transportation and other fields has surged. In the past, it was only applicable to fault tolerance in RISC platform and HP-UX environment. Products are also facing new challenges. On the other hand, companies are increasingly relying on information systems to complete critical business applications, and they are unlikely to have more professionals to perform full-time maintenance. Dual-system hot standby and clustered servers are experiencing difficulties.

NEC has launched the industry's first fault-tolerant server based on the IA architecture and supporting Microsoft Win-dows Server 2000 standard operating system environment in 2001 through cooperation with the US fault-tolerant company. NEC's Express5800/ft series achieves 99.999% reliability on Windows and Linux platforms. This real-time protection technology comes from the Fundamentals of Continuous Pro-cessing Design, which includes:

1. LOCKSTEP Technology

LOCKSTEP technology uses the same, redundant hardware components to process the same instructions at the same time. LOCKSTEP technology can maintain precise synchronization of multiple CPUs and memory, executing the same instructions in the correct same clock cycle. This technology guarantees that any errors can be detected, and even with short-lived errors, the system can resume normal operation without interruption or loss of data.

2, FAILSAFE Software

FAILSAFE software works like LOCKSTEP technology to prevent many software errors and storage and loss. The software adopts hot plug, memory mirroring, load balancing, multi-point termination failure, multi-channel I/O in Windows 2000/2003 environment, which greatly enhances the stability of continuous operation of the system. FAILSAFE manages and diagnoses feature captures, analyzes and communicates server software issues, allowing individuals to correct errors before software errors occur.

The following features of the FAILSAFE software enhance the reliability of the NEC Express5800/ft system in Windows environments: protection against short-lived hardware failures; prevention of software failure through enhanced drivers; capture, analysis and correction of software problems; The continuity of memory data is maintained; the rich error correction function can solve various errors. In order to avoid accidental failures such as physical impact, the safety fault software also provides an automatic restart function, which can save the CPU and memory data in front of the machine in real time, and avoid accidental loss of data to the utmost.

3, activation service (ACTIVE SERVICE)

Of course, if the hardware of the fault-tolerant server is permanently faulty, although the system can operate normally, the hardware must be replaced in time to maintain a fault-tolerant redundant architecture. . Fault-tolerant servers are equipped with a simple and intuitive graphical interface to manage monitoring tools (such as the NECExpress 5800/ft provides ESMPRO management software), which enables timely monitoring of hardware operation and fault conditions in the server.

The application of fault-tolerant technology has begun to enter the basic industries from the past securities, telecommunications and other fields, such as manufacturing, energy, logistics, transportation and small and medium-sized business groups and governments with 7×24 uninterrupted operational needs. NEC caters to the rapid growth of the Internet and introduces the latest stable, secure, scalable and powerful Linux version for fault-tolerant servers. The future of fault tolerance will lead to higher availability and better maintainability. According to the survey, more and more users are paying attention to TCO (total cost of ownership) rather than the initial purchase price. More companies have decided to gradually abandon the use of hot standby to maintain complex cluster servers and turn their attention to A platform with fault tolerant technology or a fault tolerant server platform.