The problem with the information technology industry is that we are fond of doing the same thing that we used to do, but giving an action a new name and telling everyone that a so-called paradigm shift has happened. Take DevOps, take big data analytics or even take cloud computing… we’ve kind of been here before with a lot of these concepts, processes and methodologies.
So could ‘bimodal IT’ be genuinely new… and what makes it happen?
What bimodal IT isn’t
In the bimodal world, IT is not broken down into mobile and desk-based technologies. Today we are far beyond that segmentation and we can quite safely work to build ‘mobile first’ (if not mobile quite a lot of the time) technologies right up to serious enterprise level systems monitoring dashboards.
What bimodal IT is
Instead then, bimodal IT is an approach where we segment technologies into two different streams or categories based upon different software application requirements, different service criticality and different maturity.
As Forbes writer Kurt Marko has noted here using Gartner’s explanation, bimodal mode #1 is traditional software with an emphasis on scalability, efficiency, safety and accuracy. So then bimodal mode #2 is software that can be defined as non-sequential with an emphasis on agility and speed.
This decoupling of software might be undertaken to a) keep legacy systems intact while the IT teams tries out new (more Agile) extensions or b) keep mission critical systems intact while the IT team tries out new cloud-centric prototype software that has yet to prove its effectiveness, scalability or ability to integrate with the mothership.
The central rationale behind cloud is scalability/flexibility, so by pushing more software into bimodal mode #2 we can make more innovation happen outside of the monolithic silos of yesteryear. At least that’s the theory anyway.
Three types of automation
Why all this discussion in the first place then? The justification for tabling the importance of this concept (whether in fact it is just a reinvention of old ideas or not) and highlighting it is the automation software. Workload automation and automated job scheduling software with automated job controls are just three of the types of auto-controls we now see increasingly deployed in the modern software application development shop.
What does software automation actually look like?
Automation in practice means management software with functions like a job library — this is a kind of directory where developers will find hundreds of production-ready templated job steps that are self-documenting. Other tools will include simulation functions — these exist so that developers and operations can simulate workflow logic and execution for testing without incurring delays imposed by actually running the job payload.
This is the market that Advanced Systems Concepts Inc. (ASCI) is going for with its ActiveBatch Version 10 product. As a piece of workload automation and automated job scheduling software the firm says it has specifically engineered its latest release to meet the demands for improved agility and bimodal deployment.
“Most IT organizations are facing unprecedented challenges due to the rate of production growth, changing business requirements, stagnating budgets and a shortage of skilled professionals,” said Jim Manias, VP of sales and marketing at Advanced Systems Concepts. “Built with an understanding of common business functions, ActiveBatch Version 10 introduces new capabilities that can cut workflow development times by as much as 50%. The benefit of these innovations allow DevOps teams to refocus on critical IT business and operational processes, thereby reducing delays, decreasing errors, and improving the speed at which tasks are released into production.”
No need to reinvent the wheel?
If products like ActiveBatch gain a foothold, then we could see software development more actively directed towards the automation functions found within.
All signs are that we now have a software industry moving towards more ‘drag and drop’ development anyway. With that in mind, it’s easy to see why functions like a library to assemble jobs and workflows faster with pre-built logic could be popular.
Why spend time researching, writing and testing code if those templates are available? Well yes, they cost money and they might not always fit the job in hand — but often they will.
According to Manias and his team, “Most organizations struggle with a combination of three to eight separate automation tools. ActiveBatch Version 10 helps reduce that number to one, due to its cross-platform architecture and jobs library. Furthermore, its predictive analytics and intelligent automation technology lets it anticipate upcoming computing needs, then assemble and later disassemble the exact mix of real, virtual and cloud resources required for the task, on the fly.”
Everybody’s trying to be more automated
ActiveBatch is not the only such automation and scheduling tool on the market. Other vendors include BMC Control-M, IBM Job Scheduler, CA’s CA Workload Automation AE and smaller interesting players like Automic — and the list goes on to Cicso and Hitcahi etc.
Automation is not a panacea; it throws up thorny issues relating to governance & control and does not fit all sizes. Nor is automation a paradigm shift in its own right, but its existence (and current popularization) is testimony to the way we are becoming more cloud-centric, more Agile-centric and more componentized and modular in terms of our approach to software.