Availability Guide for Application Design

What Is Application Availability?
Availability Guide for Application Design—525637-004
1-10
Outages Are Not Limited to the Server
Reconfiguration Outages
Reconfiguration outages are those times that the application is unavailable because of
scheduled downtime. Reconfiguration outages include those caused by:
•
Incremental reconfiguration, such as adding a disk or a communication line
•
Massive reconfiguration such as:
°
Migrating from a Complex Instruction Set Computing (CISC) system to a
Reduced Instruction Set Computing (RISC) system
°
Migrating from a D-series operating system release to the G-series, or from the
G-series to the H-Series.
°
Migrating from one version of an application to another
°
Migrating portions of an application from one platform to another
The Dynamic System Configuration (DSC) utility (on D-series systems) and the
Subsystem Control Facility (SCF) subsystem interfaces can perform most incremental
reconfiguration tasks without taking the application offline. Refer to Section 10,
Designing Applications for Change, for help in how to design applications that can be
upgraded without bringing the application down.
Outages Are Not Limited to the Server
The TNS/R and TNS/E architectures provide the most reliable computer systems
available. However, applications that run on one computer system have become the
exception. Most applications today involve computer systems from several vendors
connected by networks of various technologies.
Users of an application see only that the application is available or is not available. As
far as the user is concerned, an outage in the PC in which the client part of the
application runs means that the application is unavailable. Similarly, a broadcast storm
on the LAN to the user is an unavailable application.
For an application to be always available, the network and the client computer system,
in addition to the NonStop server, must always be available. Section 2, Overview of
Server and Network Fault Tolerance, provides information on keeping the network and
the client system always available.
Fault-tolerant hardware and fault-tolerant system software alone do not ensure a
continuously available application. The application itself also must be designed to
tolerate faults.
The NonStop Application Environment
Figure 1-1 shows the environment in which a highly available application executes on a
NonStop system. To provide continuous availability, each level requires that all levels it
depends on are continuously available. Each level must also provide its own solution to