It’s a fact that IT is filled with duplicate technology and incomplete functionality, especially where batch process management and job scheduling is concerned.
Application vendors and technology companies think they can solve this need by offering simple, date- and time-oriented job scheduling tools. We see this from Microsoft, IBM, SAP, Oracle, Infor, and many other examples. But all this does is set up a three-ring circus of unrelated technology within our IT infrastructure.
UNIX and Linux operating systems include a free feature called cron. Cron is a date- and time-oriented job scheduler with a UNIX command line interface and abbreviations for scheduling that you’d have to be a programmer or technical administrator to implement.
IT today demands solutions that can be audited and reported on for regulations and management needs. Cron has very little visibility with its command line interface. Sure, you can write a script to dump out the contents of what all the codes are—and this is great fun for the administrator to interpret—but it is about solutions not guess what this code means.
Cron is an island technology; it works for one instance of UNIX/Linux at a time and consequently cannot help with cross-platform dependencies unless you create a script that does the work. Your IT infrastructure will eventually outgrow this limitation.
Native IBM i Scheduler
IBM i has its own free scheduler called WRKJOBSCDE (work job schedule). Like cron, it is for just one instance of IBM i and has a command interface. However, it is a smidge more intuitive than cron because you can create jobs to run without being a programmer.
The bigger issue, though, is that the scheduling is still pretty basic. It lacks the ability to process dependencies, so you are constantly setting up new jobs and writing scripts to fulfill dependency processing.
IBM Advanced Job Scheduler for i
IBM i also has the Advanced Job Scheduler. Even though you must pay to use this tool, it lacks the forecasting and reporting capabilities needed to accommodate today’s highly regulated industries.
Like cron, AJS is island technology that does not reach out to other operating systems in your environment, like Windows, UNIX, or Linux. More often than not, you need the ability to monitor files being created in these other servers as a part of your processing. So, again to solve the issue your team ends up writing code.
Windows Task Scheduler
Microsoft’s Task Scheduler has the advantage of a familiar, Windows user interface, which is certainly is easier to navigate than the command line interface. But here again the problem is that it is only for one instance of Windows at a time—and it is for Windows only. Most enterprises have more than just Windows servers to worry about.
Let’s step away from the free job schedulers on your servers to take a look at your applications. Application vendors are often under pressure from their customer install base because:
- The free technology schedulers are just not adequate
- Cron, IBM i, and Task Scheduler lack a good interface
- They don’t work well with the application
The vendor then writes a job scheduler into their application because, after all, scheduling isn’t very complicated and we can add it as a feature—and maybe even charge for it!
The EnterpriseOne embedded scheduler is called Sleeper. It is a chargeable feature and integrates seamlessly with this software, but it’s rather notorious for being unreliable. Plus, it can be tricky to get IBM i or other operating system tasks that need into these schedulers.
Infor M3 Job Scheduler
Infor has acquired many business applications over the years and each may have an embedded scheduler feature. In the IBM i world, we run into several well-known applications now owned by Infor, including M3 (Movex), LX (BPCS), XA (MAPICS), Infinium, and many more.
M3 has an embedded scheduler, but it is date- and time-oriented and cannot support sharing scheduling even across two Infor products. Many of these applications exist across disparate servers and operating systems in your environment, creating more islands of scheduling within your applications.
Unlike EnterpriseOne Sleeper, these tools are free, but there are no dependencies, no cross-platform support, and no advanced scheduling—so, not very useful for coordinating jobs across your enterprise.
The free scheduler in SAP is called SMS37, and it comes with more of the same limitations. The good news with SAP, though, is that they actually recommend another vendor called Redwood to do their scheduling.
This solution is very expensive and it is not reliable for IBM i customers. It prefers to run SAP on UNIX and does not understand IBM i very well. So, if you need to coordinate IBM i processing with SAP, you cannot do it.
These are just some of the big guys out there, but the limitations of their embedded scheduler can be applied across many other applications in your environment.
The thing is, application and technology vendors provide a basic scheduler, not a workload automation tool. Basic schedulers can only handle dates and times, no critical scheduling based on fiscal calendars and no way to handle processing your tasks in sequence or across multiple operating systems.
The world of IT is complicated enough without having to try and figure out the relationships of night processing tasks in free and supplied date- and time-oriented schedulers. Regulations, the need for SLAs, and the need to stay in control of your data center operations should be enough incentive to look for more advanced tools like Robot Schedule or Skybot Scheduler.
Believe it or not, there are better options for automating your business process flow and how it fits into your data center processing.