18
RS/6000 SP and Clustered IBM
^
pSeries Systems Handbook
RS/6000 44P Model 170 (7044-170)
A High Availability Control Workstation (HACWS) option is also available on
configurations with SP standard frames only.
For more information, refer to Chapter 7, “Control workstations” on page 195.
1.5 RS/6000 SP and AIX clustering
The new e-business environment creates the need for a computing infrastructure
comprised of servers that optimize three classes of workloads: Data transaction
servers to manage core business processes; web application servers to manage
end-user experience; and appliance servers to manage specific network
functions.
The environment must provide essential computing attributes that enable
seamless integration and management of end-to-end web infrastructure.
Computing resources in this diverse computing environment must be managed
to effectively balance complex workloads as well as deliver the required Quality
of Service (QoS), Capacity on Demand (CoD), security, scalability and
manageability. This aggregation or clustered computing, include servers,
operating systems, storage, networks, applications, racks and stacks.
The emergence of Web serving and e-business as mainstream application
workloads has dramatically increased the value and relevance of clustered
computing. Clustered computing now includes “tight clusters” (as RS/6000 SP)
for high availability and single application scalability, “loose clusters” (as dense
computing elements in racks) for appliances and web infrastructure, and “grid
clusters” (heterogeneous, distributed collection of computing resources) for
e-utilities and terascale applications. These diverse cluster solutions must deliver
the needed attributes of QoS, CoD, security, scalability, and manageability.
On the IBM AIX operating system environment, we may consider the following
different cluster computing models and their main characteristics:
Sequential and Distributed Applications (little or no application alteration is
required)
– Load Balancing Cluster
•
Scale number of users or transactions throughput through application
replication
•
Allocation of incoming transactions based on load-balancing algorithms
Figure 1-4 illustrates a load balancing cluster.
Summary of Contents for RS/6000 SP
Page 2: ......
Page 6: ...iv RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 16: ...xiv RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 48: ...24 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 100: ...76 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 182: ...158 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 218: ...194 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 284: ...260 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 388: ...364 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 436: ...412 RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 477: ...0 1 spine 0 1 0 169 53 89 pages RS 6000 SP and Clustered IBM pSeries Systems Handbook...
Page 478: ......
Page 479: ......