Resource sharing: best practices

Overview: Moving from a silo to a distributed sharing model

Novice EGO administrators may implement a “silo” model for initial testing, creating resource groups that remain quite distinct and self-contained, and configuring resource plans to exclude the ability to borrow and lend resources between consumers. By maintaining such a limited model, administrators lose the advantages of EGO’s flexibility to dynamically respond to consumer and client needs and to distribute resources effectively across a cluster. To break down the silo and move to a more distributed sharing model, you must enable lending and borrowing. This can be done by altering the resource plan in a variety of ways, and ranking consumers.

Suggested best practices for resource planning include:

  • Configuring resources that are available for sharing

  • Configuring surplus resources for lending

Note that the key to effective resource distribution includes more than configuring suitable plans and policies. You must also ensure that there are enough resources distributed to the resource group(s) in order for the lend/borrow plans to make a difference. A transition from a silo to a distributed model of resource distribution requires a combination of effective resource plan configuration and the appropriate distribution of resources to resource groups in the tree.

Configuring resources that are available for sharing

The cluster administrator may set aside a number of resources for all consumers within the consumer tree to share. These unowned resources get distributed throughout the tree according to share entitlements (share ratio).

A best practice to help you better leverage the distribution of resources available for sharing is to interpret the share ratio as a minimum share value. These minimum values are complemented by the maximum share values you must also set. A distribution policy that includes resources available for sharing should always yield actual share values between the minimum and maximum constraints. So long as this remains the primary expectation, the complexities of the actual distribution algorithm do not matter.