IBM ILOG ODM Enterprise provides client-server based deployment in which planners can interactively perform "what-if?"analysis and can trigger potentially large and complex optimization solving on a remote server. The IBM ILOG ODM Enterprise Optimization Server supports concurrent and multi-user invocations, as well as programmatically scheduled (batch) jobs. In both scenarios, it manages and monitors the optimization engines that produce decision recommendations to planners.
Massive computation for solving
The ability to create multiple scenarios, along with use of “what-if?” and sensitivity analysis to investigate different alternatives and combinations, can lead to massive computational needs for the optimization solver. Planners not only don’t want to burden their desktop machines running these lengthy computations, but they also need more powerful machines to solve optimization models faster. Utilizing the power of a dedicated IBM ILOG ODM Enterprise Optimization Server, planners can study more alternatives and therefore make more informed decisions. Furthermore, multiple users can share these costly machines.
IT compliance
Today’s corporate IT departments have to maintain many applications and therefore are always looking for applications that are easier to deploy, control and maintain. The current standard to achieve these goals relies on using application servers, such as Websphere Application Server. IBM ILOG ODM Enterprise Optimization Server can deploy on an application server, in order to provide such standard compliance. Its Web-based administration console allows monitoring of deployed and executing applications on the server, making it easier to identify possible issues.
Load balancing
By deploying the optimization server on a cluster of machines, IBM ILOG ODM Enterprise Optimization Server allows them to share computational workload as a single virtual machine. Requests initiated from different users are managed by, and distributed among, all the standalone computers. This results in balanced computational work among different nodes, improving the performance of the cluster system.
Cluster: several users have their jobs dispatched on different execution nodes.
High Availability
Also, by deploying the optimization server on a cluster of machines, IBM ILOG ODM Enterprise Optimization Server can improve availability of mission-critical analytical decision support applications by using redundant nodes. High Availability (HA) cluster implementations use redundancy of cluster components to eliminate single points of failure. The most common size for an HA cluster is two nodes, which is the minimum requirement to provide redundancy.
Example of recommended architecture for Highly Available deployments.
