Blog > April 2020 > Schedulers, Mini-schedulers and Orchestration – What Does this Mean to DevOps?

Schedulers, Mini-schedulers and Orchestration – What Does this Mean to DevOps?

As we work with our customers around the world, we are hearing a consistent interest in expanding the scope of their DevOps practices, in many different dimensions. Some are keen to pull their mainframe development into their mainstream DevOps cycles, as Senior Product Manager Anna Murray discusses in her blog “Why You Should Include JCL Tools as Part of Your DevOps Strategy Others are focused on reducing the degree of isolated islands that scheduling many DevOps products in the toolchain introduces. The goal here is to avoid taking a step backwards into greater complexity (which more than one CTO I worked for said “always equates to cost!”) and fragmentation of control (inevitably creating errors and rework). 

So, where do the domains of schedulers, mini-schedulers, and DevOps orchestration all intersect? And how can they be harmonized to magnify the benefits to the enterprise?

Schedulers and the later workload automation (WLA) products are the general-purpose mainstays for getting stuff done in the data center and have been for decades. The former originally – and in some cases still – focused solely on the efficient execution of a stack of batch jobs queued for execution – processing to quickly dispatch all the jobs in the queue from beginning to end. Frequently, they are operating system- or platform-specific, such a cron for the *nix platforms or Task Scheduler on Windows. Workload automation represented the necessary evolution beyond that beginning, supporting more complex workflow mapping, dependencies and conditionality, and cross-platform coordination. Still, the primary focus has been getting a batch cycle’s work done as quickly and efficiently as possible – i.e. achieving operational excellence. 

What is a Mini-scheduler?

Some readings are now asking “So, what is this ‘mini-scheduler’ thing of which you speak?” Mini-scheduler is a term coined by one of our architects, Clay Crowell, to describing the scheduling capabilities that may be built into certain packaged or open-source technology, which has limited scope of utility. Jenkins has a sort of mini-scheduling capability built into it, that focuses on the specifics necessary for managing the continuous integration/continuous delivery (CI/CD) pipeline. However, it does not have the scope to schedule and manage the execution of production workloads – so, it is a mini-scheduler.

All the forgoing have, in their design and use, focused principally on operational execution, regardless of whether the scope is limited (execution of a single Jenkins automated build and test pipe) or enterprise-wide (completion of a vast complex interdependent nightly batch cycle at a global third-party transaction processor). They are well-tuned for executing volumes of like work efficiently.

However, they are less well-tuned to the needs of managing the front-end development activity of the DevOps cycle, coordinating the many, many discrete, rapidly evolving and ever-changing tools in the DevOps toolchain. This is where orchestration comes in, providing productized integrations with these many tools, so they can be coordinated together easily to achieve the desired development outcome.  This might include, for example, triggering some work into a Jenkins mini-scheduler, obtaining the results of the Jenkins build and test and, depending on the results, triggering a rework loop back into the development phase, or releasing to a subsequent packaging phase. 

This orchestration work becomes particularly interesting when the operations aspects of DevOps are considered. In many cases, the development being done is not focused on a simple self-contained end-user application, but a large complex multi-user, multi-technology process delivering value to the enterprise and its customers – i.e., a value stream. These require all the capabilities of the workload automation in use in production, with complex conditional workflow mappings that change from release to release, that need to roll into production at the same time as the new code builds. DevOps is not complete without the automation of the operations deployment to the same degree of quality as the development orchestration.

That is why customers around the world are adopting ASG-Enterprise Orchestrator. It offers significant and expanding integrations with DevOps tools for orchestration during the development phase, capabilities for describing the complex conditionals workflow that the delivered new / revised technology must follow, and the proven high-volume, high reliability WLA capabilities needed for operational excellence – all from a single platform. Working from a single seamless package reduces complexity (which, to CTOs, equals cost) in its operation and reduces errors (which also equal cost) in hand-offs from development to operations.

Contact your ASG representative to learn more of how you can accelerate your time to delivery, so your enterprise benefits from your development work sooner, with fewer errors and higher operational efficiency, with ASG-Enterprise Orchestrator.