The NovaScale R422 combines high-level performance and optimum cost efficiency in an ultra-dense packaging

Bull has announced a new server to its NovaScale Universal range, which combines an outstanding computing density and space efficiency, an eco-friendly energy efficiency and an efficient cluster administration.

Philippe Miltin, Vice-President of Bull's Products and Systems Division, said: “By combining unmatched computing power with the flexibility of industry standards, Bull’s solutions offer high performance computing a new world of performance, flexibility and openness. Bull draws on its expertise as a manufacturer of some of the world’s largest supercomputers to design all types of infrastructures, finally making HPC affordable and flexible. Proof that our strategy meets customer’s demands is that the University of Hanover has already chosen to deploy a large cluster based on this new range of Bull NovaScale servers, while the product was still in its development phase!”

Outstanding computing density and space efficiency
With two servers in 1U, the NovaScale R422 is a standard server with the density of blades. It delivers savings in terms of chassis-related costs, power supply, cabling and rack costs. With double the computing power within the same space, customers can do more with less space, thus significantly reducing their real-estate or rental costs, and ultimately their total cost of ownership (TCO). More than 5.4 Teraflops peak performance can be packed within a single rack, delivering maximum performance per square meter.

Eco-friendly energy efficiency
The NovaScale R422 chassis features a highly efficient power supply unit (PSU) that allows significant savings on energy costs on the long term, thus considerably lowering the TCO. Whereas conventional PSU efficiency rarely reaches above 75%, the NovaScale R422 offers an outstanding power utilization rate of 92%.

Efficient cluster administration
The NovaScale R422 clusters are delivered with a choice of the Bull Advanced Server (BAS) Linux environment or Microsoft’s Windows Compute Cluster Server 2003. The BAS software suite for Xeon-based servers is built on top of standard Red Hat Enterprise Linux 4 or 5. This integrated and comprehensive package – designed by Bull – includes the best of Open Source and industry-standard software, together with specific features developed by Bull for high-performance computing (HPC) clusters. It includes comprehensive management tools to allow an entire cluster – including the platform and interconnect network – to be as easily monitored and managed as a single system.

Technical specifications
The NovaScale R422 features up to two dual or quad-core Intel® Xeon® processors (5100 or 5300 series). This rackmount server incorporates innovative technologies including the ‘Atoka‘ motherboard jointly designed by Supermicro and Intel, that allows two servers to be fitted into a single 1U chassis: i.e. up to 16 cores, or up to more than 170 Gigaflops in 1U.
The NovaScale R422 is ideally suited for use as a compute node in NovaScale clusters, in combination with NovaScale R440/460 service nodes. With its on-board InfiniBand connector, it offers unlimited scalability while minimizing interconnect cost. Whether for a small, medium or large-scale configuration, anything up to several thousand NovaScale R422 servers can by clustered together to deliver the required computational power.

The NovaScale R422 is available now.

More information about NovaScale servers

High-availability, an ever-increasing requirement
Major public and private sector organizations have a natural interest in the various techniques available to improve the productivity of their IT assets. The higher the level of real availability, the better the return on investment (ROI) on IT solutions. Increasing computing power is not the only way to drive up productivity.
The high-availability (HA) market is expanding rapidly. In the previous decade, HA services were mainly associated with large-scale systems running ‘mission-critical’ applications for the enterprise. But today, HA is key to four times as many systems: supporting applications that underpin daily business activities, such as production overflows and round-the-clock mass-market e-commerce online transactions. The ever-increasing globalization of business activities is also driving up demand, with the need for servers to remain accessible around the clock from any point on the globe, and no matter what time zone the user may be in.
Another factor driving up demand is the growth and development of system architectures and infrastructures. HA services are now possible on most configurations and operating systems, no matter what size they are.

Quality assurance, guaranteed results and financial terms meeting customer’s demand
When they first came onto the market, HA services were undertaken as standalone projects in the same way as an application development might have been: and this is still the case for many service providers. This meant that a number of specialists had to get involved in the early stages of the project, and significant specialist system supervision and task-execution resources had to be put in place. All this came at a cost.
Bull realized from an early stage that these kinds of high added-value services would eventually become a necessity for most of its customers, and so established a range of pre-defined services around a core support contract: HA999 (High Availability 99.9%). Located at Echirolles near Grenoble, our specialist European support center has been up and running since 1999; monitoring, diagnosing and carrying out tasks remotely our customers’ mission-critical systems. Moving away from ‘project mode’ towards a more structured, professional and automated approach to support has significantly reduced operational loading and resulted in a big increase in the number of systems now being supported under HA contracts.

Professional HA999 services for end-to-end support
The High-Availability Center of expertise at Echirolles in France is the cornerstone of Bull’s HA initiatives. All our customers with systems managed under HA contract are monitored by this specialist center; with all customer requests, automatic alerts and queries being handled directly by this group. All hardware and software configurations are archived and monitored by the center. We have developed several dedicated tools, including automated 24/7 alert tools that have considerably reduced the number of straight system failures, thanks to proactive actions. When local Bull offices get requests for on-site interventions, these are always monitored by the HA Center right through to their operational conclusion. Every quarter, the customer receives a comprehensive report of all the tasks carried out on each system, along with both an analysis of the problems encountered and appropriate recommendations. Serving as a reference point from one quarter to the next, this report enables both the Bull team and the customer’s own staff to improve the way each system is managed under an HA contract.

Services offered by the Center include:
Configuration audit. The first stage generally involves producing a ‘status report on the current system’, as well as a technical report covering all known versions of hardware, firmware and software installed at the production site. Recommendations for any further development work are also given, to ensure overall technical coherence. At this point, tools are also installed that will collect and upload data to the European Support Center.
Optimizing version management. Analysis of the existing versions can also be accompanied by a report on all the technical and organizational actions to be implemented, with the aim of:
• Improving the level of consistency between versions of hardware, firmware and the operating system, so that it is compatible with the various versions of applications
• Controlling the attendant risks and costs of a change in system maintenance
• Securing the changes and system updates, from the production point of view
• Optimizing and synchronizing the evolution and updating of versions, to achieve better operating reliability and better configuration control.
Proactive surveillance of in-production versions. Bull’s European Support Center automatically collects version control data for installed systems, and analyses this is on a month-to-month basis to:
• Ensure systems running under the operating environment concerned conform to an authorized standard for versions
• Enable on-going verification of the integrity of versions installed and in use
• Help anticipate future needs for developments and technical upgrades.
Production continuity support. This service is designed to support the production team, particularly during off-peak periods, by checking the conditions for production system continuity, and helping to restart production following an incident. This support service is provided in the form of remote surveillance of the availability levels of IT production resources by our specialists at the European Support Center, who are familiar with the relevant operational management details and the customer’s disaster recovery plan. The customer keeps overall technical responsibility for system operation.

Therefore with its European Support Center, Bull has developed a highly structured, professional approach to delivering the High-Availability contracts entrusted to it by customers, with:
• Proactive support (systematic measurement, change management, ITIL practices…)
• Automatic technical alerts
• A dedicated, high-availability support center running 24/7
• Monitoring, diagnostic and remote intervention tools
• Skilled local teams of technical staff ready to take action rapidly on site
• Close technical collaboration between customer teams and Bull.