In-Depth

News in Brief

HP pay-for-use computing; updated IBM Director;l MS Datacenter Server; VisualAge C++ for Linux

• By Stephen Swoyer
• 03/13/2003

HP Announces Pay-for-Use Computing Offering

Hewlett-Packard Co. this week unveiled a new utility computing technology that can measure the rate of utilization on each CPU in its Superdome servers.

As a result of its new offering, HP says, it can charge customers for their actual usage on a monthly basis.

HP’s new offering is the latest in its Utility Data Center initiative, which is similar to the E-business On Demand and N1 utility computing models touted by IBM Corp. and Sun Microsystems Inc., respectively.

Utility computing is a concept that describes the virtualization of computing resources—such as storage, servers and networks—to increase performance, lower costs, and enhance manageability.

True to form, HP says that organizations that leverage its new pay-per-use offering can more quickly respond to upticks in seasonal activity or to unexpected business opportunities and only pay for the processing power that they need, when they need it. HP points out that pay-for-use can benefit customers during dry spells, as well, as—again—they only have to pay for the computer power that they actually use.

HP’s new pay-for-use offering will be delivered through its HP Financial Services group. It is installed on-site, and works by recording the actual utilization of each CPU. This data is automatically collected, encrypted and securely transmitted to HP's billing engine.

HP says that its new pay-per-use offering is available in North America, Europe, Africa and the Middle East. It is expected to launch in Asia-Pacific shortly.

IBM Updates Director

IBM Corp. this week announced a revamped version of IBM Director, software that it has developed to manage large deployments of Intel-based servers. Big Blue says that version 4.1 of IBM Director incorporates more than 20 performance and intelligence enhancements.

IBM’s Director supports all of Big Blue’s xSeries Intel-based servers, from server blades to its 16-way x440 server to Integrated xSeries Servers (IXS) running on its iSeries minicomputers. IBM Director automates common tasks—such as inventory, monitoring and alerting, event actions and system health checks—and facilitates integration in heterogeneous environments. It also includes autonomic features that, Big Blue says, maximize server availability and reduce downtime.

New in IBM Director 4.1 is an extensible framework that supports plug-ins to facilitate advanced management functions. The new version of Director also features configuration wizards for quick deployment and set-up.

In addition, IBM ships a variety of management modules for Director 4.1. The upshot, says Big Blue, is that Director 4.1 can provide a single point of management for most tasks. Director 4.1 modules include a Server Plus Pack, which features predictive tools that help to optimize server performance and high availability; Remote Deployment Manager, which provides drag and dropfeatures to remotely replicate the installation of multiple systems; Software Distribution Premium Edition, a tool that facilitates remote software distribution; and Application Workload Manager, which increases server utilization by protecting the availability and performance of workloads on a server.IBM Director 4.1 will be available worldwide March 31 for all of IBM’s Intel-based servers and is also available for a fee on other industry Intel-based servers.

Windows Datacenter Program Overhauled by Scott Bekker(courtesy of ENTMag.com)

Microsoft's Windows Datacenter Server has been about more than technology features since its launch in late 2000. Fittingly, the tightly controlled support program that makes Datacenter Server so different from other Microsoft server operating systems is getting a complete overhaul for the Windows Server 2003 launch. Microsoft unveiled details of the program last month, including the new name, the "Windows Datacenter High Availability Program."

When Windows Server 2003 ships on April 24, several changes will go into effect with the Windows Datacenter Program:

• The program will be called the "Windows Datacenter High Availability Program."

• The process for OEMs to resolve customer problems and get quick fix engineering work from Microsoft on Datacenter Server, known as the Joint Support Queue, is being changed into a more regimented program known as the High Availability Resolution Queue.

• Requirements that OEMs recertify their Datacenter Server-based systems in a 14-day stress test will be waived for more minor component changes. In those cases, a special 1-day stress test will be allowed for requalification.

• The support program is being broadened beyond the system OEMs to include straight service providers.

• Microsoft's support organization will also step in and begin offering direct service for Datacenter.

• New services are being built into the program, including pre-installation assessments of customers.

• For customers who don't have a need for high availability on their Datacenter Server-based systems, Microsoft will begin allowing customers to use basic support instances and credit card payment support for problems with Datacenter Server.

Read the complete text of this article here: http://www.entmag.com/news/article.asp?EditorialsID=5725

Big Blue Delivers VisualAge C++ for Linux on pSeries

IBM last week announced version 6.0 of VisualAge C++ for Linux on pSeries, which is based on its VisualAge C++ development environment for AIX.

VisualAge C++ 6.0 for Linux on pSeries supports SuSE’s Linux Enterprise Server 8 for pSeries running on Big Blue’s Power3 and Power4 architectures. Its C compiler supports the latest ISO C 1999 standard, while its C++ compiler adheres to the latest ISO C++ 1998 standard.

The new development environment includes IBM's compiler optimization technology, which enhances support for 32-bit and 64-bit optimization in both compilers. In addition, VisualAge C++ 6.0 for Linux on pSeries generates optimized code for the Power3 and Power4 architectures.

Finally, Big Blue says that its compilers can exploit shared-memory parallel processing through automatic and explicit parallelization.