Task distribution
The section describes how multiple parallel tasks from Control Tower are routed and distributed among multiple Bots (RPA Nodes).
There are two (or more) different types of tasks or business processes. For example for two lines of businesses to automate different types of applications. And there is a need to divide the nodes into two separate pools. So for one type of task to use the first pool of nodes, and for the second type of task to use the second pool.
Two high level approaches are described below. Usage of one or combination of these will allow to configure distribution and segregation strategy of any complexity.
Approach 1 - Segregate hubs into two separate quotas
You can configure Bot Manager to have two (or more) quotas, each quota to have its own list of hubs. Then each individual quota is dedicated to a specific business processes or tasks (or development teams).
For the configuration guide refer to Bot Manager.
- Configuration is centralized on the Application server (where Bot Manager resides)
- Configuration is done by system administrators (ssh access to the APP server is required)
- Business processes need to specify the quota to utilize
- Only hubs can be specified in the quota files, not individual nodes
- Each new hub needs to be added to one of the quotas upon creation
The approach is more suitable for isolated pools of nodes. Some reasons for the strict segregation could be: regulatory or security requirements, business processes nature differ significantly (in terms of applications used) and development teams of the business processes are segregated (no direct communication and collaboration path, e.g. may be the case when implementations are done by different service providers).
Approach 2 - Configure RPA nodes capabilities
caution
Not recommended for production usage — no guarantee of full RPA cluster utilization.
You can specify for each node the custom capabilities it has (usually the applications it has installed on), and setup your bot configs to specify what capabilities from nodes executing them are required. For the configuration guide, refer to RPA Nodes Capabilities.
- Configuration is required on every hub or node of what capabilities it has.
- In the business process logic, the RPA steps should specify what capabilities it requires.
- Fine-grained control of each individual node is possible with such approach.
- Design and initial configuration of the capabilities as a rule is done by development team(s) with subsequent business process deployment documentation created as part of operational guide for higher environments.
The approach is more suitable for the case when nodes may be shared for multiple types of tasks. The approach gives more control to solution designers and business process developers on the decision of effective topology utilization.