Abstract :
Summary form only given. Today´s applications are being built in new ways. The advent of service-oriented architectures, modern networking protocols and industry-standard components at all levels open seemingly endless possibilities for new applications and services. This article focuses on the particular challenges for servers and how these challenges are being addressed. The NonStop Enterprise Division of HP (formerly Tandem computers) has built massively parallel, hardware and software fault-tolerant servers. These servers power many of the most challenging transaction processing environments that exist, including major stock exchanges, funds transfer applications, credit and debit card applications, E911 applications, major telephony and instant messaging applications, and integrated hospital applications. The system is based on a message-oriented architecture and thus bears many similarities internally to the architectural principals being espoused for today´s service-oriented architectures. HP is bringing to market a new generation of integrity nonstop servers, based on an innovative design using Intel Itanium 2 processors and leveraging many other industry standards. The article describes how the new generation of servers can accomplish even higher levels of availability than current systems while at the same time leveraging industry standards and meeting other extreme requirements. The article touches on the lessons learned over the years when dealing with mission-critical applications and how those lessons apply to today´s distributed, service-oriented architectures. The underlying message is that careful use of encapsulation, limiting the use of distributed algorithms to carefully controlled situations, and overall design rigor are required even more than ever
Keywords :
distributed processing; network servers; open systems; dependable server; distributed architecture; message-oriented architecture; mission-critical application; service-oriented architecture; software fault-tolerant server; Application software; Concurrent computing; Fault tolerance; Hardware; Hospitals; Network servers; Protocols; Service oriented architecture; Stock markets; Telephony;