HewlettPackard will partner with Cisco Systems and Oracle Corp. to develop products for higher-performance Unix servers — products that promise to keep an enterprise network up and running 99,9 percent of the time — from the hardware to the operating system to the database to the application. Conversely, Hp is implementing a similar solution across its NetServer line running Windows NT and cluster server software from Microsoft.
In the collaboration with Cisco and Oracle, the foundation for the new program, called 5nines:5minutes, will be new servers based on the HP-UX operating system. With production expected to begin during this year’s first half, the new systems will run the IA-64 microprocessor, and will be upgradeable via a hardware module to the upcoming Merced microprocessor when it becomes available sometime in late 1999.
True 99.9-percent uptimes, or no more than five-minutes of downtime, won’t happen until the year 2000, said Bill Russell, VP and GM of the enterprise systems group at HP. “We will have 4nines availability in 1999, with thc release of a major new system.” tie said by then, the new systems will have even more hot-swappable capabilities, from the hardware to the operating system components. Also on thc hardware side, Cisco will focus on three levels of network redundancy: the physical, thc logical link and the backbone protocol. Products as a result could be hot-swappable linecards and hot standby routing protocols, said Jayshree Ullal, VP of marketing for the enterprise line business at Cisco. Ill the area of software, Cisco will work to deliver Web-based management tools, and improve upon current software tools. Meanwhile Oracle, will continue to improve upon its Parallel Server technology, which is part of the Oracle8 network computing database,, in the areas of fail-over protection. The promise of a network that is up and running virtually all the time is a difficult one to keep. To meet this challenge, the companies will address certain categories of potential failures. Thc first is avoidance, to prevent downtimes in the first place. HP will use its consulting services to help customers build highly available and integrated infrastructures by reviewing current information technology environments for points of potential failure and implementing appropriate architectures.
The second category, regeneration, will address network recovery, or how quickly a network can refuel itself without disturbing applications. This requires quick, transparent repair of any component in the IT infrastructure, via a hot-swappable model. The third category, inclusion, includes extending uptimes across both hardware and operating systems. Simplicity, the’ fourth category, for eliminating human error and reducing installation complexities. This will be done through software that can manage, monitor and administer an entire information technology environment from a single console; in this case, through HP’s OpenView product. The goal will be to monitor an enterprise from one Web browser user interface that contains drag-and-drop functions within the database management tools for, for example, watching for traffic slowdowns on a network. HP says this new program will change the way vendors will compete. Said Mr. Russell: “Today, suppliers differentiate themselves through products that offer system availability. But as businesses are forced to compete globally, and as the Internet becomes a more common way for companies to do business, availability will be measured in terms of the customers’ perception of the availability.”