For anyone tasked with implementing Opalis, there are a number of solution-level questions that will immediately burst into your head. 

You will have a bunch of questions along the lines of:

  • How many servers will I need?
  • How many concurrent jobs can I run?
  • How long will the jobs take to complete?
  • What specs will my server need?

I liken these questions to someone who wants to build a house:

  • How many rooms should my house have?
  • How many people can live in my house?
  • How long will it take to build my house?
  • How big should my house be?

What are the correct answers to these questions?  The reality is that the answer is that most common IT answer - “It Depends”.

There are some things that you can run through a calculator and get an answer.  IT Process automation is not one of them :)

You will however be able to categorise them into groups and apply some average-based calculations.  But how do you categorise them?

The very first thing you need to do, is the non-technology stuff.  If you have not documented your current manual processes and understood what you are trying to achieve, then you have no point of reference to work out the answers to those architectural questions above.

This is also the time to ensure that the processes are efficient, and indeed actually necessary.  If you automate an inefficient process, you will not gain the full benefits of the automation.  You will just have a faster, inefficient process.

Once you have this complete, then you will be able to test with a selection of the processes to get a feel for the resources required and the performance parameters for your artefacts.

As with any technological implementation, a number of parameters come into play to define the performance of Opalis.  The more you spend, the more hardware capacity you have and the quicker your automations should run (I use the word ‘should’ because external influences like the systems you integrate with, network, firewalls and the like will play a role).

So you will need to test with a few different hardware types to get a feel for what performance gains you will get as a result of the additional power.  There will be a point where additional hardware has no or little effect.

So, did I answer the question?  No :) because there is no correct answer.  But if you invest in the up front Analyst time, do your testing and document what you have and what you want, the answers will become clear to you.