Fujitsu 200/200R, Брошюра и характеристики

Страница: 7 / 8

High Availability Cluster Systems
When your business cannot afford to stop, even for a short time, you need more than just
high reliability in each server. Such mission critical requirements are met by GP7000F's
clustering system. Multi-processor clusters linked to shared disk sub-systems minimize
service interruption when accidents or failures occur. Such clusters also provide
continued service when one server fails, or when components are off-line for routine
maintenance.
Range of solutions
The GP7000F's Clustering Systems offer a range of highly
available applications and environments in addition to the
basic durable and available system. The GP7000F's solutions
include highly available databases and transaction
processing systems, as well as network and shared printer
systems.
Availability
The hardware components are duplicated to maximize
system availability. This prevents total system failure or
data loss when a single component fails. In the event of a
potential server failure, the operational system can be
quickly transferred to a standby node to avoid service
disruption. The IP address, the operating node, MAC
address, and node name can also be transferred.
Hot standby technology
GP7000F can also be used in a "hot standby" configuration.
In this case, the standby node is in a hot standby condition
and is connected to the same shared disk system as the
operational node. This is different from the standard way of
transferring data and applications following a failure of an
operating node. With the GP7000F's "hot standby"
configuration, the time taken to restart the job is shortened
dramatically.
Operation
GP7000F clustering solutions remove the difficulties of
installation and operation that have previously restricted the
use of clustered solutions. Excellent system configuration
tools, such as the cluster installation wizard, automatic
hardware search and configuration function, WWW browser-
based operational setup, and centralized control function
take away the complexities of cluster management. In
addition, these tools also greatly improve the levels of system
security in the operational system.
Scalability
A GP7000F clustering system not only provides improved
availability when data or processes are transferred due to
node failure, but it also dramatically improves overall system
performance through parallel processing.
Maximum 128CPUs
Scalable up to a maximum of 16 nodes/128CPUs by clustering.
Auto isolation
When a node fails, it is automatically separated from the cluster.
It can then be serviced offline. The remaining nodes in the
cluster reschedule the work and continue to operate without
service interruption.
High-speed interconnect system
Fujitsu's Synfinity high-speed interconnect technology is used
to transfer information between clustered nodes. Synfinity is the
world's fastest interconnect technology and provides a data
transfer rate of 240MB/sec. in both directions. This high-speed
transfer significantly improves the parallel database processing
performance of Fujitsu's SymfoWARE and Oracle's Parallel
Server systems.
Cluster Operation Control Screen
Cluster System Configuration
Shared Disk
Failed
Application Application
Hot Standby
Business
Clustering
System
Cluster
System base
Printing NFS Network Database
Tran-
saction
Parallel
Database
Processing
Cluster basic
S/W
Disk
multiplex
path
Volume
Control
Highly
Reliable
File
System
High
Speed
Interconnect
(VIA)
LAN
multiplex
path
Availability Scalability