Tackle TCO Head On.
PC servers have become a crucial element of all enterprise networks. Companies increasingly rely on servers to run critical business software and provide continuous availability of key corporate data to remote locations worldwide. As businesses depend on continual data access, server maintenance and repair teams must be available 24 hours a day, seven days per week. And, while corporate IT staffs are shrinking, network managers' responsibilities are expanding. These factors present a challenge for companies looking to maximize the productivity of their enterprise networks while minimizing overhead expenses and support costs.
Thanks to recent advances in hardware and software, PC servers can now integrate seamlessly into an existing IT structure. This provides cost-saving functionality that lowers a company's total cost of ownership (TCO) while easing the growing maintenance burdens of network managers.
LEVERAGING HARDWARE ADVANCES KEY TO EFFECTIVE SERVER MANAGEMENT
Monitoring key hardware is the first step in preventative server maintenance. Today, leading-edge servers incorporate advanced monitoring systems capable of proactive system management of the vital functions affecting the health of your server. Key areas to monitor are CPU temperature, system temperature, power supply voltage and temperature, critical fan errors, ECC memory, and hard drive status. Some servers can detect changes that may impact future operating capabilities and alert the network administrator to potential hazards.
When a network server has a memory error, it can affect hundreds, if not thousands, of users on a network. Memory-intensive software applications can tax system memory and require larger and larger memory devices. Memory protection has advanced and is available today in two forms: parity memory and error correction and code (ECC) memory. While some vendors continue to rely on the older parity memory technology, this technology can only detect odd numbers of bit errors and cannot locate or correct the erroneous data. Therefore, if parity memory detects an error because there is no inherent ability to correct the error, the server will halt operation. In contrast, the ECC memory scheme is the better choice because it improves error detection capabilities and provides the added benefit of being able to correct certain errors. With proactive server management capabilities, the server will alert the IS staff in the case of a memory change.
The most costly component of network management is the personnel required to maintain the network itself. Centralized system management means improved efficiency, as administrators are free to perform management duties remotely. New technologies keep the network administrator in touch with remote servers, even in the unlikely event of a critical system failure. Capable of transmitting an alert even if the server goes down, the Emergency Management Port's (EMP's) out of-band access feature allows the network administrator to perform basic server functions remotely.
Another key component of industry-leading server hardware is the [I.sup.2]C Management Bus. A standalone data bus used to transport communications signals, the [I.sup.2]C Management Bus acts as a miniature network on the motherboard. This standardized method for communication of motherboard-specific data enables quicker, more efficient data transmission, rerouting communication signals across its specialized "information highway" in order to optimize performance.
Compliance with the SCSI Accessed Fault-Tolerant Enclosures (SAF-TE) specification, an industry initiative defining chassis health management status, is another important aspect of server management. The SAF-TE specification was designed for extendibility, flexibility, and openness, as well as to increase networking availability and improve productivity of the LAN administrator. Adherence to the SAF-TE specification facilitates a more complete and automatic monitoring and alert notification of the storage subsystem--both locally and remotely. SAF-TE allows the server to communicate its status by generating an industry-standard SNMP alert, which is then intercepted by the management software application. In this way, issues such as a power supply failure or drive failure can be automatically paged to the network administrator or maintenance provider via a single software agent.
NEW SOFTWARE TECHNOLOGIES ROUND OUT THE HARDWARE MANAGEMENT EQUATION
Hardware server management represents just one-half of the total manageability equation. Efficient server management also requires a user-friendly software interface that presents data in a clear, easy-to-use format. To complement the product's hardware functionality, look for a server solution that offers compatibility with leading management software solutions, such as Intel's LANDesk products, Computer Associates' Uni-Center TNG, Hewlett-Packard's Open-View, and Tivoli's TME 10.
For example, LANDesk Server Manager (LDSM) is bundled with many leading server solutions and is an open, cost-effective server management tool. LDSM allows network administrators to communicate the operational status of server hardware, minimize the likelihood of server downtime, restore servers more quickly when problems occur, increase reliability and availability, manage servers remotely, and tune servers for optimal performance. LDSM delivers a multitude of emergency management recovery tools, providing network administrators with comprehensive reactive management capability. It also provides a proactive management solution, since it performs ongoing data collection and real-time monitoring of up to 250 parameters of server hardware and performance.
Many servers today are also optimized to work with enterprise management software such as CA's Unicenter TNG. Designed to simplify and enhance server management in today's distributed enterprise environments, Unicenter TNG offers an effective platform for server administration and management.
Unicenter TNG Framework integrates the management of IT resources via a common interface and infrastructure, simplifying the task of managing complex, networked environments. It lets network managers monitor IT resources at any time, from any location, using a Java-based Web browser. Supporting more than 40 operating systems, it is the only enterprise management framework to provide out-of-box support for a broad range of platforms. Its Real World Interface, available in 2-D, 3-D, or Web-based format, lets you navigate intuitively through your entire enterprise network. Unicenter TNG Framework also provides the foundation for managing applications in the area of security, operations, networks, databases, help desk, and storage.
Until very recently, system-level information could not be ported efficiently to an enterprise management software solution. Advances in server design now enable true integration of both the system and enterprise layer, ensuring that system information can be viewed by leading enterprise software applications. Leading server providers typically bundle specifically engineered plug-ins with their systems to facilitate seamless interaction between software platforms.
Advances in server technology are providing a powerful foundation for reducing ownership costs across the enterprise. As companies depend on servers as a hub of communication to customers, suppliers, and employees, IT managers need to look for well-rounded server solutions that tackle TCO head-on, delivering new functionality that enhances manageability.
Circle 251 for more information from Toshiba
Greenlund is senior director, server product marketing, Toshiba America Information Systems, Computer Systems Division, Irvine, Calif.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Technology Information; new server management technologies|
|Date:||Jan 1, 1999|
|Previous Article:||Reduced ownership costs key to competitiveness.|
|Next Article:||Securing a smooth transition for the future.|