Scalability refers to the ability of the environment to meet the needs of an increasing number of users and external services in a way that is predictable in terms of performance.
The main scalability consideration is whether the cluster architecture is dynamic and additional servers can be added to increase the number of users (simultaneous or not). You should also consider whether other components of the cluster can be added to the design later. It may be simple to expand the size of the shared repository because database servers are usually made to scale, but it may be more complex to change load balancers or implement redundant load balancers.
There is often a relationship between scalability and high availability. If the system works under normal circumstances but stops functioning under load, this may violate your availability requirements. You should also consider whether availability requirements will change along with increased user load. For example, as the cluster-based BI solution is rolled out to more and more users, high-availability of the cluster becomes more critical. In this case, you may need to add redundant load-balancers or other component upgrades to ensure future availability. Your initial cluster design should recognize and allow for this expansion.
Once you have defined your scalability requirements in detail, use this information when sizing your cluster.
Recommended Comments
There are no comments to display.