Architecture/Technology Overview

ZKLWHýSDSHU
7
+3ý 1HW6HUYHUý /;Uýåèíí
Server Clustering To further assure system availability, server clustering is often
employed. HP supports four different programs from several vendors including Microsoft
Cluster Server, Vinca Co-Standby
Server, NSI Software Double-Take
®
and Novell SFT III,
all for HP NetServers. Depending upon the cluster program, either 2-node or 4-node
clustering is possible, increasing system uptime for the data center.
Remote Management
The HP LXr 8500 includes several powerful management features built into the architecture
of the server. Micro-controllers and I
2
C buses permit system event log management.
Remote management is available via the RMC, a plug-in card.
Industrial Design
The physical design of the HP NetServer LXr 8500 has also been optimized for the data
center in several ways.
The server is rack-mounted in a 7U (12.25 in. or 31.12 cm) rack height, “rackable and
stackable”, which is important as the data center consolidates servers. Since this is an
8-way server, the processor density is very high, over a very small footprint. The design
conserves floor space, which is at a premium as the amount of equipment continues to
increase. Five servers can be stacked in one 2-meter tall rack (41U) to a maximum
40 Pentium III Xeon 550 MHz processors per rack. Tremendous processing power in a very
small space.
High processor density produces high levels of heat. The fan cooling system has been
optimized to remove the heat load, with extra fan capacity to spare. (The importance of
redundancy was explained earlier.)
Since availability is so important in a data center, this server has been designed for quick
replacement of key components. Identification of component status is aided by easy-to-
read LED indicators for the overall system, fans, power supplies and PCI cards, each of
which can be replaced without shutting down the system. Adding more memory or more
processors requires that the system be shut down for a very short time.