Professional Documents
Culture Documents
Best Practices For Workload Automation Ebook PDF
Best Practices For Workload Automation Ebook PDF
Workload Automation in
the Bimodal Era
sponsored by
Dan Sullivan
Best Practices for Workload Automation in the Bimodal Era
iii i
Best Practices for Workload Automation in the Bimodal Era
iiiiii ii
Best Practices for Workload Automation in the Bimodal Era
Copyright Statement
© 2016 Realtime Publishers. All rights reserved. This site contains materials that have
been created, developed, or commissioned by, and published with the permission of,
Realtime Publishers (the “Materials”) and this site and any such Materials are protected
by international copyright and trademark laws.
THE MATERIALS ARE PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND,
EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE,
TITLE AND NON-INFRINGEMENT. The Materials are subject to change without notice
and do not represent a commitment on the part of Realtime Publishers its web site
sponsors. In no event shall Realtime Publishers or its web site sponsors be held liable for
technical or editorial errors or omissions contained in the Materials, including without
limitation, for any direct, indirect, incidental, special, exemplary or consequential
damages whatsoever resulting from the use of any information contained in the Materials.
The Materials (including but not limited to the text, images, audio, and/or video) may not
be copied, reproduced, republished, uploaded, posted, transmitted, or distributed in any
way, in whole or in part, except that one copy may be downloaded for your personal, non-
commercial use on a single computer. In connection with such use, you may not modify
or obscure any copyright or other proprietary notice.
The Materials may contain trademarks, services marks and logos that are the property of
third parties. You are not permitted to use these trademarks, services marks or logos
without prior written consent of such third parties.
Realtime Publishers and the Realtime Publishers logo are registered in the US Patent &
Trademark Office. All other product or service names are the property of their respective
owners.
If you have any questions about these terms, or if you would like information about
licensing materials from Realtime Publishers, please contact us via e-mail at
info@realtimepublishers.com.
iiiiiiiii iii
Best Practices for Workload Automation in the Bimodal Era
111 1
Best Practices for Workload Automation in the Bimodal Era
The Challenges
IT managers and executives face two constant challenges:
• How to acquire and develop staff with the skills needed to operate a dynamic IT
department
• How to centralize and consolidate IT operations to maintain a cost-effective
operation that delivers on the needs of the business
At first glance, these may seem like unrelated problems, but in fact, they are more
intertwined than first appearance suggests. This interconnection will become clear as we
delve into the details of each. At the same time, it will start to become apparent how the
role of workload automation can help address both of these challenges.
IT Resource Gap
IT professionals with in-demand skills have email inboxes littered with messages from
recruiters. Subject lines with phrases like “Immediate Need,” “Phone Interview Hire,” “Top
Salary for Sys Admins with knowledge of x,” where x is any of a long list of in-demand skills,
such as virtualization, cloud administration, and information security. This practice of
borderline-spamming of IT professionals is the canary in the coal mine facing many IT
organizations: they do not have the IT skills they need and the employees with top-valued
skills are sought after by competitors, collaborators, and other businesses alike.
• 60% of survey respondents are concerned about the impact of delays in filling open
positions
• 25% have suffered a loss in revenue due to unfilled positions
• 41% of respondents report increased workloads are driving down morale
• Customer service is adversely impacted in 30% of cases
• 25% of respondents report an increase in mistakes and lower quality work product
222 2
Best Practices for Workload Automation in the Bimodal Era
These statistics speak volumes about the difficult situations facing organizations and
employees, but one additional statistic stands out: 22% of respondents report high
turnover because of the added demands on staff due to unfilled vacancies.
Thus, the staff members whom organizations depend on to meet business needs and make
up for unfilled positions are increasingly at risk of leaving due to frustration, low morale,
and excessive workloads (see Figure 1.1). Given the fact that so many organizations are
facing similar skill gaps in their IT operations, those with skills in high demand will find
they likely have a range of options when looking for a new position. What was a problem
filling new positions is now compounded by the fact that businesses need to retain the
skilled staff they have in the face of stiff competition for those skills.
Figure 1.1: Fewer employees have the skills in demand to support IT operations,
creating a skills gap.
333 3
Best Practices for Workload Automation in the Bimodal Era
Figure 1.2: An IT resources gap is growing as hardware, software, and services are
increasingly available at relatively low costs while the growth in skills needed to
fully utilize those tools lags behind (Source: http://info.advsyscon.com/it-
automation-blog/addressing-the-growing-it-resources-gap).
Although it might take a team of well-trained software developers with years of experience
to make a useful tool, such as a business intelligence reporting tool, that same tool can be
applied by thousands of developers in businesses around the globe. Those developers must
be familiar with business intelligence reporting, capable of working with business
professionals to discover requirements, and have an understanding of state-of-the-art
methodologies, such as Agile, to deliver and sustain a finished product. Those skills are in
demand.
Leveraging Existing Staff with Automation
As the discussion so far has illustrated, executives and managers are facing difficult
decisions about how to allocate their staff resources to projects and operations. If they take
experienced IT professionals away from operational duties to develop new applications
and services, they risk disrupting operations by pulling experienced administrators from
crucial day-to-day operational tasks.
One way to deal with this challenge is to automate administrative tasks, such as running
regularly scheduled jobs. This kind of workload automation is commonplace in business.
Systems are backed up, data is exported from transaction processing systems and loaded
into data warehouses, inventory is analyzed, and new orders placed. These are the kinds of
operations that are well-structured business processes and amenable to automation.
444 4
Best Practices for Workload Automation in the Bimodal Era
They share common characteristics:
The problem with this all-too-common scenario is that the short-term technology needs are
emphasized over the long-term needs of the organization. Solving problems in isolation
leads to a complex unmanageable environment consisting of multiple workload
management tools, scripts, and systems (see Figure 1.3). Some are homegrown solutions
written in the favorite language of the developer who happened to be tasked with getting a
process up and running. Some of the tools are third-party applications purchased by a
department to put an end to the disparate collection of Perl, Python, and PowerShell scripts
that have accumulated in the department’s workloads. At some point, a department
manager might lay down the law and declare all processes be developed using a standard
tool—standard for that department anyway.
555 5
Best Practices for Workload Automation in the Bimodal Era
Figure 1.3: Silos of automation tools emerge when individual developers and
departments try to separately address the common problem of workload
automation.
666 6
Best Practices for Workload Automation in the Bimodal Era
The skills gap problem can be addressed, to some degree, by workload automation, but not
if the decision-making is made at the level of individual developers or even at the
departmental level. A centralized approach that takes into account broad business
requirements and architecture is essential to avoiding, or at least mitigating, the challenges
described to this point.
Figure 1.4: Without a centralized workload strategy, organizations struggle to keep
disparate solutions and tools operating together.
777 7
Best Practices for Workload Automation in the Bimodal Era
As noted earlier, workload automation can help compensate for the skills gap. Tasks that
must be performed repeatedly can be automated in a variety of ways. The flexibility of
choosing tools for workload automation is both a benefit and a detriment. It is a benefit
because individuals and teams can rapidly deploy workload automation scripts using tools
they understand and can readily support. The downside emerges as this practice continues
with different individuals or teams choosing different tools.
From the perspective of a CIO, the organization has created a system that now depends on
a variety of tools requiring different skills sets. The choices each team made were no doubt
rational from their perspectives. Unfortunately, those perspectives are often limited to
their own project requirements, budget constraints, and scheduling deadlines. Gartner
research found that this pattern is so prevalent that it is not uncommon for organizations
to support three to eight different workload automation tools. If you are fortunate, your
organization supports only three workload automation tools, but some organizations are
supporting as many as eight workload automation tools. This conclusion seems out of line
with effective IT practices.
CIOs tend to take a broader view of the organization and the portfolio of services they have
to support. Thinking like a CIO with that broader perspective, you could see the advantage
of a consolidated approach to automation. Ideally, the workload automation solution will
serve the needs of multiple projects, operations, and departments. The benefits are clear:
888 8
Best Practices for Workload Automation in the Bimodal Era
Scripts
Cloud
ERP
Orchestra2on
Centralized
Infrastructure Workload Managed File
Provisioning Automa2on Transfer
Security/ Data
Compliance Warehousing
Repor2ng
Figure 1.5: Centralized workload automation enables a single point of control and
problem resolution.
The ability to support a wide range of workload scenarios is especially important today
because IT organizations are shifting their structures to support both day-to-day
operations and more Agile development and deployment of new services. An organization’s
workload automation solution should support both constituents of this bimodal structure.
999 9
Best Practices for Workload Automation in the Bimodal Era
• Reliability
• Efficiency and price optimization
• Plan-driven approach
• Waterfall methodology
• Ability to work with long-term enterprise vendors
• Month-in-scale projects
• Tend to be IT centric
In short, Mode 1 “keeps the lights on.”
Mode 2 addresses the needs of business to respond rapidly to changing market conditions
and customer opportunities. Agility is the hallmark of Mode 2 IT. It is characterized by:
101010 10
Best Practices for Workload Automation in the Bimodal Era
We might think mostly of Mode 1 operations when we focus on workload automation, but
that would be a mistake. Mode 2’s more agile approach introduces new kinds of complexity
into an organization. Mode 2 practitioners are experimenters. They may develop a new
process and modify it almost continually in attempts to optimize it. This kind of
experimenting and tuning would not be possible without workload automation tools.
Manually rerunning workload steps would be too time consuming and prone to error.
Gartner recommends that organizations deploy workload automation tools for Mode 1
operations that address the need for reliability and the ability to automate operations from
end to end. For Mode 2, the analysts recommend a tool that enables continuous
deployment. This kind of rapid deployment is feasible in Mode 2 operations but less
appropriate for Mode 1 where change management helps mitigate the risk of disrupting
operations.
Using Mode 1 and Mode 2 operation models is a way to balance competing needs within an
organization. Some tasks need to be executed in a highly controlled and reliable way.
Financial operations, such as generating ledgers and running payroll, fall into this category.
The cost of error can be so high in these areas that it is worth being cautious and slow in an
effort to minimize risk. Other areas, such as deploying a new version of a mobile
application, can tolerate higher levels of risk. If an application is rolled out to employees
and it causes problems for 10% of the users, a fix can be created and deployed rapidly. Yes,
there may be some slight disruption in the efficiency for those 10%, but the risk is worth it
because it is part of a larger innovative process that, over the long term, leads to more
efficient operations.
111111 11
Best Practices for Workload Automation in the Bimodal Era
Consider examples of the impact of disrupted IT services. On one day in July 2015, three
major service disruptions occurred:
• United Airlines experienced a router issue that caused flights to be suspended for 2
hours. By the time the incident was resolved, 800 flights had been delayed and 60
flights had been cancelled.
• The New York Stock Exchange, the largest in the world, shut down for more than 3
hours. The incident was related to a software update that was rolled out. It turned
out that some customer gateways were not properly configured for the new
software.
• The Wall Street Journal website was down for about 1 hour at the same time the
New York Stock Exchange experienced its outage.
These three incidents are not causally related, but the fact that they all happened on the
same day highlights the potential for service disruptions across a range of industries. Some
disruptions are relatively minor, such as a website being down for a short period of time,
while others are costly events that will attract the attention of C-level executives.
Given that service disruptions should be expected, the question becomes how should we as
IT professionals plan and prepare for them. We cannot anticipate all possible problems and
their root causes, but we can frame processes and workload management procedures with
an understanding that they need to be resilient to failure. Broadly speaking, we can
approach this challenge with manual procedures or with automated methods.
121212 12
Best Practices for Workload Automation in the Bimodal Era
Another benefit of workload automation solutions is that they can help address the IT skills
gap. By deploying a workload automation tool, organizations can develop a set of skills
across the organization that will serve operations spanning the business.
Summary
IT operations are crucial elements of business operations. Some operations must be
reliable and highly available while others are subject to near continual change in response
to new business opportunities. IT departments across industries face a critical skills gap
that leave us putting a premium on solutions that can be reused throughout the
organization. Workload automation tools and best practices can help address these
organizational needs.
131313 13
Best Practices for Workload Automation in the Bimodal Era
141414 14
Best Practices for Workload Automation in the Bimodal Era
• Time-based processing
• Relatively stable set of jobs
• Reliable custom scripts
Closing the books and running payroll are the types of operations that run on a regular
schedule. Essentially, there is no unpredictability in these tasks. Fixed, time-based
processing was typical with early batch jobs. This setup was a significant advantage to
application developers and system operators. The tasks that were run last week or last
month would be run in the same order this week or this month. There was no need to
account for complex logic that would trigger the execution of some processes and not
others.
The time-based processing that characterized early batch jobs meant that businesses could
depend on a fairly stable set of jobs. Only so many business operations were automated,
and they tended to address core business operations, such as managing revenues and
payments. These types of business operations are stable and do not change much. In fact,
they do not change much today thanks to well-defined and agreed-upon accounting
procedures.
With a relatively stable set of jobs comes a stable set of scripts to run those jobs. After
programmers write and debug the code behind a batch job, it can continue to run for years
without change—as long as business requirements do not change. There may be some
minor changes in the business logic of an operation—for example, a change in payroll law
that requires a change in the way the process performs a calculation—but these do not
necessitate large restructuring of code or development of new batch jobs.
This description of batch processing sounds almost mythical in the way it describes a time
when simplicity was a defining characteristic of business operations. Of course, developers
had to deal with complex issues, but they centered more around working with scarce
computational resources and the lack of modern development tools. Managing workloads,
once developed and debugged, however, did not entail the challenges IT professionals face
today.
151515 15
Best Practices for Workload Automation in the Bimodal Era
If we think about what must be going on behind the scenes of these operations, we can
pretty quickly see that the characteristics of batch processing no longer hold (see Figure
2.1). In particular, these new operations are not:
161616 16
Best Practices for Workload Automation in the Bimodal Era
Figure 2.1: Time-based batch processing is too limited for today’s workload
management requirements, which must be capable of responding to a wide array of
event triggers and file constraints.
171717 17
Best Practices for Workload Automation in the Bimodal Era
In addition to this, one of the biggest challenges IT managers face is finding and retaining
skilled staff. IT positions go unfilled and business requirements go unfulfilled. Existing staff
members are asked to continually do more with less resources to meet demands. One of
the needs that can get overlooked is the development of custom scripts to perform a new
set of business tasks in a controlled, repeatable, and dependable way.
In those cases where an individual develops custom scripts, those individuals are typically
also the ones who have to maintain them. When developers are expected to write scripts
on short schedules to meet a very specific requirement, it is no surprise to find the
resulting script is highly tailored to the existing requirements. As new requirements
emerge, the scripts must be modified. If the developer did not have time to anticipate the
types of changes that might be needed in the future, their scripts may be difficult to adapt
to the new requirements. In fact, it is not uncommon in such cases to simply start from
scratch and write a new script to meet the new requirement. This cycle creates a
proliferation of scripts that IT developers must continually maintain.
In the past, when businesses ran a limited set of batch jobs, there was time to develop and
debug code that one could reasonably expect to run for long periods of time. After
programmers wrote a batch script, the primary focus turned to making sure it ran as
expected and addressing problems as they arose (e.g., running out of storage space for an
unusually large set of data). Today, a developer might write a script and then turn her
attention to writing another script or to debugging one that a colleague wrote several
months ago that no longer works because of changes in an upstream process that generates
input for the now problematic script.
At this point, there is an unsustainable set of IT practices that can no longer meet the needs
of businesses. This argument is not a criticism of past practices but a recognition that
business demands on IT are changing. Custom scripting and batch processing is the best
practice when an organization needs to implement a small set of relatively stable,
repeatable processes. As the need for the number of custom scripts grows to meet
emerging business needs, the demands on IT grow in unsustainable ways. IT departments
must change the way they meet workload needs to leverage their staff skills, minimize
maintenance, and ensure reliability of operations. This reality creates the need for
workload automation.
181818 18
Best Practices for Workload Automation in the Bimodal Era
191919 19
Best Practices for Workload Automation in the Bimodal Era
Figure 2.2: Example library component for performing secure file transfer.
202020 20
Best Practices for Workload Automation in the Bimodal Era
For example, a component for specifying an input file could accept a filename as a
parameter and perform a consistent set of operations, such as:
Figure 2.3: Example of a content-rich workflow designer.
212121 21
Best Practices for Workload Automation in the Bimodal Era
Reduces the Need for Custom Error Handling and Alerting Logic
In addition to streamlining the development of individual processes, workload automation
tools help businesses manage collections of these processes (see Figure 2.4). In particular,
by using a common set of error-handling and alerting logic, system operators can use a
single, centralized reporting and alerting tool to monitor the state of multiple workloads.
Businesses that depend on custom scripts can find themselves developing individualized
procedures to differentiate the error logging and alerting of those scripts.
Figure 2.4: Dashboards in workload automation can provide a consolidated view of
the state of multiple workloads.
222222 22
Best Practices for Workload Automation in the Bimodal Era
This section highlights some of the most important features to look for when selecting a
workload automation tool:
232323 23
Best Practices for Workload Automation in the Bimodal Era
• As noted earlier, time-based batch processing can address the needs of a small set of
workload automation requirements. Event-driven automation is an essential feature
of modern workload automation. Common events can trigger workload execution:
o Receipt of an email
o A message delivered in a message queue
o System startup or shutdown
o Web service execution
o Database trigger
Robust workload automation tools have mechanisms to detect and respond to
events across the IT ecosystem. Also look for automation tools that can balance
workloads. You might find, for example, that some servers are running at capacity
while others are underutilized. The ability to move jobs among resources can help
optimize workload execution (Figure 2.5).
Figure 2.5: Workload automation tools should allow administrators to move jobs
across server resources to execute workloads efficiently.
• The contents of the workload automation library strongly influence the overall value of
the platform. Libraries should include tools for common functions:
o Managing files and directories
o Transferring files between servers
o Logging workload information
o Integrating with databases, enterprise resource planning (ERP), customer
relationship management (CRM) and other enterprise applications
o Working with legacy scripts
o Invoking API functions
242424 24
Best Practices for Workload Automation in the Bimodal Era
252525 25
Best Practices for Workload Automation in the Bimodal Era
IT managers will appreciate the consolidated reporting, alerting, and management that
comes with a workload automation tool. This functionality is especially useful when self-
service portals and mobile device access are provided. Perhaps the biggest advantage of
using workload automation platforms is that they provide the ability to deploy new
business processes faster. As with any software tool, it is the combination of quality
software and best practices that help businesses realize the most benefit from their
investment.
262626 26
Best Practices for Workload Automation in the Bimodal Era
Summary
This chapter has examined the composition of a workload automation solution with a focus
on framing workload automation in terms of the history of batch processing. Businesses
will realize the optimal benefits of automation by choosing a workload automation tool that
supports key requirements and combines that platform with a series of best practices.
272727 27
Best Practices for Workload Automation in the Bimodal Era
282828 28
Best Practices for Workload Automation in the Bimodal Era
When businesses consolidate and centralize workload automation, they often see several
benefits:
292929 29
Best Practices for Workload Automation in the Bimodal Era
This scenario is repeated over and over and over again in IT shops that use many different
tools for automating their environments.
Perl Ruby
Developer Developer
PowerShell
Developer
Figure 3.1: Developers have overlapping skills and can sometimes work across
application frameworks but doing so often requires deep knowledge of a
programming language or tool to solve complex operational problems.
303030 30
Best Practices for Workload Automation in the Bimodal Era
Consolidated Reporting
We are living in a world where human attention has become a scarce commodity. Never
before in history have we had so much information at our fingertips.
Despite this availability of data, businesses are struggling with ways to derive meaningful
insight from various sources. As more and more data becomes available for analysis, IT is
dealing with an abundance of applications, databases, and systems to manage, in addition
to the data dependencies between these systems.
• Some custom scripts write logs to the server on which the workload job is running
• Other custom scripts write to a consolidated logging system on the network
• Each custom script has its own pattern for logging different levels of information
• Log messages have a variety of structures that each requires custom code to parse
Now imagine a manager has asked for a status report that shows the results of a logical unit
of business work, such as fulfilling an order from the time it is entered into a web
application to the time it is shipped and delivered to the customer. One piece of the process
writes data to a custom log format from the web application to indicate the status of the
original order. The inventory system logs data about changes in inventory levels with
respect to the order. The shipping workflow interacts with a third-party service to schedule
delivery using the lowest cost carrier. Each of these pieces is written in a different
programming language using a different log format.
313131 31
Best Practices for Workload Automation in the Bimodal Era
323232 32
Best Practices for Workload Automation in the Bimodal Era
333333 33
Best Practices for Workload Automation in the Bimodal Era
These questions arise with any business software. Tracking licenses and optimizing the use
of licenses can be challenging. Different vendors have different licensing models. Some may
have per-user licenses while others have server-based licensing. Add the additional
considerations of running multiple virtual machines on a single server and the move
toward utility computing in the cloud and you begin to see the hidden overhead of using
multiple software tools for the same task.
Patching and Maintaining Workload Software
All software must be maintained. Even code we do not write ourselves still has some level
of maintenance overhead. This requirement usually means we set up tasks to check for
updates and apply those updates on a regular basis. If we are using a piece of software
throughout the organization, it is almost certainly worth the effort to standardize and
automate the patching process. When departments use specialized tools, such as different
automation tools, there may be a tendency to rely on the application administrator to
remember to manually update software on a regular schedule. Needless to say, the latter
strategy is prone to human error that can lead to unpatched software.
Training Staff on Software
Another problem with multiple applications is that they require multiple people to be
trained to use and maintain them. This problem is the same we have identified in previous
chapters. IT staff have limited time to develop skills and perform their jobs. The more
software we add to the enterprise portfolio, the more stress we put on IT personnel to keep
that software running efficiently. This, in turn, requires them to invest time to learn the
new software and understand how it integrates with existing applications, programs, and
devices.
One of the best ways to reduce the burden around managing software licenses, patching
and maintaining software, and training IT staff is to decommission unnecessary software.
Why should your organization deploy several software packages when one would suffice?
It is true that standardizing on one tool will require some migration effort. Scripts written
for decommissioned tools will need to be reimplemented in the consolidated workload
automation platform, but that is a one-time cost. The long-term benefits of a consolidated
approach will likely outweigh the short-term costs of reimplementing workload scripts.
343434 34
Best Practices for Workload Automation in the Bimodal Era
• Reduced need for diverse and difficult-to-maintain skills for different workload
automation tools
• Consolidated reporting
• Reduced dependence on custom scripting
• The ability to decommission unnecessary software
The benefits do not end with these obvious personnel and financial benefits. There are
additional benefits that stem from increased ability to deploy workload automation
services more quickly and reliably to your organization.
• Libraries of tested code that allow for reuse, rather than rewriting
• Workflow application programming interfaces (APIs)
• Self-documenting operations
• Visual scripting tools
Each of these features can contribute significantly to the ability to build and maintain an
array of workload automation processes throughout your organization.
353535 35
Best Practices for Workload Automation in the Bimodal Era
Workload automation platforms provide libraries of tested code to perform these kinds of
operations and more. They are especially important for saving developer time. Not only do
in-house developers not need to write these scripts from scratch, they do not have to
maintain them. Developers can simply reuse these libraries of tested code, instead of
rewriting scripts over and over again. Also, workload automation tool vendors can invest
time to research the best ways to implement a function, such as authenticating users and
processes. This can lead to a more efficient and secure implementation than might be
possible if an in-house developer is under time constraints to get a functioning custom
script done quickly.
Workflow APIs
Workload automation solutions that support API accessibility promote more connected,
flexible environments. With API accessibility, organizations can incorporate methods and
properties of any technology or application into the automation solution’s content library.
As a result, organizations are virtually unlimited in scope: any technology with an API can
be easily incorporated into workflows, allowing developers to reuse rather than rewrite
scripts using pre-built Job Steps in the content library. As more organizations adopt
service-oriented architectures and the use of microservices, workload automation will fit
with that design model if APIs are readily available.
Self-Documenting Operations
Documentation is a necessary and valuable part of any software. Good documentation is
not always easy to come by. A developer on a tight schedule can get a script up and running
without writing documentation, but she cannot avoid writing necessary code. If forced to
choose between writing code to implement a feature and writing documentation describing
how the implementation works, most developers will choose the former.
Workload automation systems can help mitigate this problem by employing self-
documenting operations. This function can be employed through the implementation of
libraries, modules, and functions that use naming conventions for functions and
parameters that make clear the purpose of the function.
363636 36
Best Practices for Workload Automation in the Bimodal Era
Figure 3.2: Visual scripting tools help increase the speed in which workload scripts
are created.
Visual scripting tools can help speed development by supporting drag-and-drop access to
libraries and APIs, context-sensitive highlighting, and visual display of information that can
help developers debug workload processes.
373737 37
Best Practices for Workload Automation in the Bimodal Era
• Running short jobs to check whether another job has finished and resubmitting
themselves to keep checking manually
• Starting a complex workload process but leaving it idle while waiting for the
preceding job to finish
• Having a human operator monitor processes and manually start workload processes
as needed
Modern workload automation tools allow organizations to employ resources only when all
job dependencies are satisfied. They also leverage advantages of cloud computing and the
associated “pay as you go” model that is typical with cloud computing.
In addition to these obvious advantages, consolidated workloads provide for centralized
management and reporting. This functionality can help provide insights into operations
and allow systems administrators and application developers to find ways to improve
overall performance and optimize multiple workload processes (see Figure 3.3).
Figure 3.3: Consolidated workload automation reporting tools can help systems
administrators and application administrators optimize performance of multiple
workload processes.
383838 38
Best Practices for Workload Automation in the Bimodal Era
393939 39
Best Practices for Workload Automation in the Bimodal Era
Be sure to inventory custom scripts as well. Doing so can be more difficult because there
may not be a central, up-to-date repository of information about custom scripts. A two-
pronged approach is suggested. First, work with application owners to create a
consolidated list of known workload scripts that are running to support ongoing
operations. Second, use scripts to scan servers for cron jobs or other batch-scheduled jobs
running on each server. In an ideal world, the list of workload scripts known by application
owners and systems administrators will match the list generated by the server scan. In
practice, there will likely be differences.
In addition to creating an inventory of workload scripts and tools, be sure to inventory the
skills available in your staff. Also consider what you expect to need in terms of skills and
assess the skills gap, if any. Plan training, if needed, to align with the deployment of the new
workload automation system.
This point in the process is a good time to think about the architecture of your future
workload automation tool. Think in terms of a consolidated infrastructure. Here is a chance
to eliminate silos of workload automation and custom scripts.
Finally, develop a project plan for migrating existing scripts to the new platform. Consider
the risk associated with any disruption in service. It often makes sense to migrate low-risk,
low-complexity workloads first. This approach allows your staff to ensure the platform is
installed and configured correctly without risking a high-impact disruption.
Summary
Centralized workload automation through consolidation and coordination can bring
substantial benefits to an organization. These stem from several factors, including the
benefits of consolidating and centralizing, improved speeds for building and deploying the
workload, optimization of resources, and IT burden reduction. As you plan to deploy a
workload automation tool, be sure to consider the next steps outline provided in this
chapter to help ensure a smooth and efficient transition to your new consolidated
workload automation platform.
404040 40