When businesses start small and have little capital to invest, they tend to keep their computer systems simple, typically using a single database server and a storage array to support a database. However, as a business grows, this simple database configuration often cannot handle the increased workload. At this point, businesses typically upgrade their hardware by adding more CPUs and other required resources or they add servers to create a cluster. Increasing resources, or vertical scaling,g is like placing a temporary bandage on the system—it solves the current situation of increased workload and resource demand but does not address the possibility of future workload increase. Instead of adding more memory or CPUs to the existing configuration, organizations can add servers and configure them to function as a cluster to provide load balancing, workload distribution, and availability—a process known as horizontal scaling.

The growth potential for a vertically scalable system is limited because it reaches a point at which the addition of resources does not provide proportionally improved results. In contrast, horizontal scalability enabled by clustered hardware can distribute user workload among multiple servers, or nodes. These nodes may be relatively small and can use inexpensive standards-based hardware, offering economical upgradeability options that can enhance a single large system. In addition, clusters offer both horizontal and vertical scalability, providing further investment protection.

Download