The application is considered highly available as it approaches the 99.9999% of availability commonly referred to as the five 9s of availability. Availability of a system is the combination of availability of both components of solution, the infrastructure and application, such as:
- The hardware layer, such as servers and storage
- The operating system, such as Linux or Windows
- The base services, such as the database
- The applications, such as Exchange Server
HIGH AVAILABILITY ACHIEVED
Availability can be improved by eliminating single points of failure (SPOFs). There is a conceptual limit on this, and the data center on which the servers are located is a SPOF. To alleviate this, there are a number of different solutions available for Disaster Recovery. Disaster recovery refers to the capability to recover an entire data center to a different location.
It is easier to build a high availability solution as early as possible; starting from the design phase and using standard components are key for building a cost effective solution. Availability can be improved on an existing installation by minimising single points of failure.
One method to improve the hardware availability is to implement clustering by using two or more computers or nodes for a common set of tasks. If one computer fails, then the others can take over the failed service. A second computer can be used as the redundant backup for the first computer. This design supports high availability.
CLUSTER TECHNOLOGIES
A cluster can be defined as a collection of interconnected complete computers or nodes, that appear on a network as a single unit.
The cluster is managed as a single system or operating entity. It is designed to:
- Tolerate component failures.
- Support the addition or subtraction of components in a way that is transparent to users.
- Accumulate resources for processing power.
Clustering becomes an important concept for both high availability and disaster recovery discussions.
There are two basics approaches for clustering:
- Hardware clustering: This requires specialised hardware. It usually involves a strong investment, both on the technical resources, for example, with special training and maintenance cost.
- Software clustering: This uses standard hardware and is a cost-effective solution for high availability.