Sample deployment scenarios

The following deployment scenarios illustrate some basic configurations taken from actual case studies. Your organization might look similar to one of the examples below, depending on the size of your network, the various bandwidth restrictions between clusters and the number of relays and servers. The main constraint is not CPU power, but bandwidth.

Pay careful attention to the relay distribution in each scenario. Relays provide a dramatic improvement in bandwidth and should be thoughtfully deployed, especially in those situations with low-speed communications. Relays are generally most efficient in fairly flat hierarchies. A top-level relay directly eases the pressure on the server, and a layer under that helps to distribute the load. However, hierarchies greater than two tiers deep might be counterproductive and must be carefully deployed. Multiple tiers are generally only necessary when you have more than 50 relays. In such a case, the top tier relays would be deployed on dedicated servers that would service from 50-200 second-tier relays. The following examples help you deploy the most efficient network layout.

Note that additional servers can also add robustness to a network, by spreading the load and supplying redundancy. Using redundant servers allows failback and failover to be automated, providing minimal data loss, even in serious circumstances.

With the correct deployment of servers and relays, networks of any size can be accommodated. Beyond the examples shown here, your IBM support technician can help you with other configurations.