Infrastructure requirements for RPA
Infrastructure-wise, there are 2 types of RPA solutions on the market today
- On premise RPA
- Cloud RPA
Infrastructure requirements for both On-premise RPA
RPA environment setup should be based on organization’s processes and policies. Infrastructure setup are very strict and tightly controlled by different groups. Setup different environments (Dev, Staging, Prod).:
- Application access on servers – If the setup is done on VM server, application capabilities should be tested.
- Security Policy – VM security policies should be looked at and make sure that it will not interfere in application and RPA tool functionality.
- Role based access control for all environments
- Tools for support: All RPA deployments use external applications like like OCR, word, excel, outlook, etc. Ensure that your setup accommodates them in all environments
- Supporting application server versions – Some applications may have different version of desktop and server. If infrastructure is setup on VM servers, this point also should be kept in mind.
Infrastructure requirements for both On-premise RPA and Cloud RPA
- Licensing of RPA tool – Licensing terms are different for each tool vendor so this needs to be discussed and appropriate licenses should be acquired for Dev and Test
- RPA tool support for Virtual Environment – RPA tools needs to be evaluated for VM support.
- Application update policies – IT team should own all RPA platforms and technology. All patches, security updates and upgrades should consider existing RPA bots ability to interact.
Guidelines to create your environment
- Initiate it soon: Infrastructure setup should not be an afterthought. It should start while working on POC. This should start early in the lifecycle. Generally, infrastructure setup SLAs are longer and multiple levels of approval are needed. Early start on setup and design will help the project a lot.
- Create environments: Ensure 3 distinct environments such as dev, test and production. This is a one-time activity and has to be done perfectly. It protects your existing running automations from issues.
- Bot setup: Dev, test and prod should have at least one bot each. Dev and test should replicate production’s setup; that is, it should have the same software versions, screen resolutions, etc.
- Branch Merging: We must always merge code from the dev branch to test branch and then to the production branch. Ideally no code changes should occur in test and dev.
- Monitoring and infrastructure maintenance: RPA is not meant to run 24×7. If it is meant to be 24×7 then bots should take over from the other bot. Operators must schedule system restarts and other maintenance activities in a systematic manner.