KE Capacity Planning

I see where there has been some recent posts around the Knowledge Exchange product.  The excerpted information below provides some good guidelines for evaluating KE usage in a given deployment.  It discusses the major components that impact performance.

The versatility of Knowledge Exchange ensures that every customer’s deployment is unique. As such, there are no hard and fast rules to predict when an additional KE server should be added to support the needs of the enterprise. There are, however, some guidelines and variables to consider when evaluating KE usage in any given customer deployment.

The major elements that most impact performance and, therefore, the number of users that can be supported on a given Knowledge Exchange server include:

                        Server Configuration;

                        Type of User; and

                        Usage Patterns.

Server Configuration

Any Knowledge Exchange server installation should minimally adhere to the base requirements outlined in the KE installation guide. Based on the other variables outlined below, the KE server may benefit from additional memory and storage. If the KE server is not dedicated and is supporting other applications and users, this can also impact the number of KE users that can be supported with adequate performance.

Configuration variables that impact usage include:

                        Number of CPU’s;

                        Amount of Memory;

                        Storage Size; and

                        Dedicated or Non-Dedicated to KE.

Type of User

There are generally (2) primary KE user classifications: rich client and viewer. A rich client is characterized as a power user. These users typically are doing modeling as a core component of their daily work and are frequently checking in and checking out objects from KE. Viewers, in contrast, are accessing Knowledge Exchange through their web browsers and leveraging KE’s simple editing and collaboration capabilities. They may be reviewing models housed in KE and annotating them, for example. The profile of the KE user community will be a factor in the overall number of users a given KE server can effectively support.

Usage Patterns

In addition to the type of user, the usage patterns of the user community should also be considered when planning for an additional KE server.

Usage pattern factors include:

                         Location of users;

                         Concurrency of user access;

                         Check in / check out patterns;

                         Time of day

                         Number of objects

                         Size of objects

                         Size of notebooks;

                         Number of workflow models

                         Number of activities

                         Number of custom objects.

One example Knowledge Exchange implementation supports (50) Rich Clients and (500) Viewers. However, each customer’s implementation is unique. All of the variables noted above play a role in determining the number of users an individual Knowledge Exchange server will effectively support.

Thanks

Gerry Victoria

Metastorm