Clustering and Redundancy

A key technique to maximize server uptime is clustering two or more servers together. Usually, this process also utilizes redundancy (i.e. adding multiple identical components) to eliminate single points of failure, thereby maximizing the reliability of a hosted solution. Loosely defined, a server cluster is made up of a set of connected servers, working together for a common purpose. As a result, it can be viewed as a single system. Individual components within a cluster are connected to each other through fast local area networks (LANs). Compared to a single server, clusters improve availability and performance and are more cost-effective than a single server that provides comparable availability and speed. 

High Availability Clustering with Redundancy

High-availability clusters are also known as HA clusters or failover clusters. HA clusters incorporate redundancy into the system architecture to eliminate single points of failure, usually by synchronizing the use of multiple data facilities and network connections. High availability clusters are especially valuable when hosting essential services such as e-commerce websites, file sharing, important databases, and customer service and internal business applications.

Redundant servers in groups or clusters continue running even when individual system components within the cluster fail. HA clustering detects failures in software/hardware and rapidly restarts the application on another system or node without requiring administrative intervention. This automated process is also known as failover. Without clustering and failover, if a particular application crashes on a server, that application and service will not be available until the server crash is fixed.

Clustering with Load Balancing

When high performance is needed to manage heavy loads, an additional clustering technique known as load balancing is used. This is where cluster-nodes or multiple servers share their computing power to provide better overall performance. This arrangement still provides all the benefits of redundancy and failover that come with clustering. Contact us today to further discuss our customized options.