Divide and Be Conquered — Tearing Down Silos for Modern Mainframe Excellence
This article is a written adoption of a presentation I first gave at SHARE Columbus 2022 titled “IT Careers: The Fall of the House of Silos”. I hope that this is a more thoughtful approach to that conversation.
Whether you’ve been in IT for long or not, you’ve likely heard of the concept of siloing. Chances are that you stopped to read this article because you either don’t like the idea of siloing or just can’t stand yet another someone speaking out against them. For the folks who aren’t as up-to-speed as you are, let’s take a quick pit stop for a brief definition.
Siloing is the isolation of a system, process, or department from others. As it pertains to most of us in the tech field, it typically means the security folks are the only ones doing security work, the networking folks are the only ones doing networking work, and the folks in storage are the only ones performing the storage functions. If you’re haunting the mainframe (specifically z/OS) space, you likely exist in a silo in some capacity.
At first glance, the concept of siloing should resonate with most of us who have visited a doctor: when we have foot problems, we might be referred to a specialist known as a podiatrist. Our knee-jerk response is empathetic towards silos, after all we are social animals. As social beings, we tend to be drawn to commonalities. It’s not far-fetched to think that the security experts flock to their security-minded ilk because they can lament and empathize over how foolish us lowly engineers are for trying to use triple DES in 2022.
Much like the social side of siloing, practically silos have served us quite well. In the early days of computing, concepts like security, networking, and storage were far too new and far too complex to be generally understood by anyone but those charged with their keeping. It didn’t make sense to expect the people protecting datasets and user IDs to know the ins and outs of VTAM and SNA. Complex topics, especially in their infancy, seem to require pioneers to specialize in them before the rest of us get a gist of what’s going on.
While silos have historically served a functional purpose in IT, the rallying cry for their dismantling has been heard for quite some time now. So, why the influx of hating on silos? Specifically, in mainframe administration, they just take a whole bunch of time. In the best case, rigid silos can prolong a task (like installing a piece of software) from taking a week to taking several weeks. In the worst case, silos can dissuade a would-be administrator from staying in the industry. Imagine this: That software installation (again, especially on z/OS) typically involves provisioning storage, defining started tasks, creating security profiles, editing PARMLIB members, and, in the case of most modern solutions, connectivity to other nodes, either in the cloud or on the local network. In a strictly siloed environment, finding these people and getting their (valuable) time can take what seems like forever.
As digital transformation continues to blaze ahead as the buzzword du jour, we’re seeing more topics and technologies become entangled in commonalities and codependence. No longer does the mainframe sit as an isolated piece of technology in the enterprise landscape. Now, to meet the needs of the business, the mainframe has to chat with distributed systems and even cloud infrastructure and services. Extending the silo paradigm to Linux and Windows systems and cloud services off-platform, one can easily see how the number of required experts would grow exponentially.
Now that we know that silos might not provide the agility we need to accomplish business objectives in a timeframe reasonable for our customers, what’s the solution? I would assert that the heroes we need to save our bacon are what most people would call “generalists”. The generalist is often relegated to the often partially quoted, “a jack of all trades is a master of none,” as a way of saying that generalists just don’t know enough about one thing to be useful. While I tend to agree that knowing how to start 50 recipes is not as useful as knowing how to finish one, the rest of the quote goes on, “[…], but oftentimes better than a master of one,” to allude that knowing a lot of things “well enough” can be occasionally more useful than just knowing one thing very well. To continue our cooking analogy: I can make spaghetti, garlic bread, and a simple Italian salad to a satisfactory level (just ask my family). This combination can make for a complete, carb-loaded meal, but if I could only make garlic bread extremely well then I suspect only our 3-year-old would be happy at the table. I don’t need to be a master chef in each dish to make a meal, which is perfectly sufficient in over 99% of the times I’d like to feed the family.
So here’s where my call to action comes in: become a generalist! A properly trained generalist can set up their own environments, implement basic security profiles, make configuration changes, and install software without much expert intervention. Generalists serve the enterprise well as they can nimbly traverse teams as needs ebb and flow. These heroes of the modern world are more effective change implementers as they have a more comprehensive awareness of the impact of a change. How often have you heard the storage team explain to you how the database folks just keep asking for more space without knowing the work involved? With a competency in other disciplines, those pain points are eased with empathy and awareness making a more cohesive strategy.
But beware, while the generalist offers unparalleled utility, it’s important to call out a few things to watch out for:
- The siren song of a strong generalist makes it easy to rely on them in all times of need, leading to exhaustion. Being aware of your or your generalists workload is critical to avoid burnout.
- The obvious lack of depth means your generalists cannot perform at a specialist level across all disciplines. Maintain mindfulness of who your specialists are in each discipline and have enough of them ready for critical situations.
- With a buffet of disciplines to learn at once, it can take some additional time for a new generalist to be productive. Successful strategies I’ve seen often involve a tour around the different disciplines with each stop involving becoming an introductory member of that team and performing basic responsibilities. This tour often ends with the individual selecting, or being placed on an as-needed basis, with one of those disciplines to develop a deeper skillset.
So now that I’ve (hopefully) sold you on the benefits of being a generalist, how do you get there? As a mentor, I often prescribe mentees to be observers in the early phases of their careers to develop a breadth of awareness and skills. When you ask for help from someone experienced, don’t simply request a task be done. Ask for a moment to watch that task be completed, write down your questions, and ask them at an appropriate time. While most of the information will likely pass you by, these gaps can be augmented through official learning channels, building professional relationships internally and externally, and independent research.
This idea of being proficient enough in multiple subjects is a concept called “T-shaped skills”. Also known as being a generalizing specialist, this path really helps one find their passion while being useful across the disciplines. The trick to success is identifying that oh-so-important “Good Enough” line. To me, that line is where an individual can complete 90% of the tasks required for that discipline, but still relies on a specialist for that final 10%. It might feel like that’s a tall order, but through careful monitoring you will hopefully find that a majority of your disciplines asks are simpler tasks that just take time, conversation, and human eyes. This proficiency can be built upon through classification of incoming work and distributing work appropriately to the available resources who would not only complete that work but benefit from doing so.
Allow me to be clear, employing this strategy does not eliminate the need for a specialist. The foundation of this approach requires enough specialists (or education) to not only build skills, but handle any critical situations. Specialists, while integral to cross-training, reviewing, and handling the most advanced tasks of a discipline, are empowered to focus on maintaining their expertise and pursuing other interests and individual growth.
So where does that leave you? Well, to be able to put a fine point on this depends where you stand in your career:
- If you’re early in your career, I would advise you to continue building your breadth of skills. Gain as much exposure to the enterprise as you can so you can find something that excites you. Your career is unfolding ahead of you and it will be long. It is paramount to find something you enjoy and can continually grow in doing. Finally, focus on building relationships both internal and external. Some of the most helpful people are simply experts in their field looking to share some knowledge.
- For my late career colleagues, a trend I have been noticing is that the most valued individuals at an organization are those that share their knowledge. As an established expert, you have decades of wisdom that both early and mid-tenure professionals can learn from. You are key components to the sustainability of not only the environments you work in, but the industry. Finally, as important as it is for our early tenure professionals to do what they love, it is just as important that you continue following your passions through the sunset of your career.
Siloes have served and will continue to serve many industries well, but as the demand for agility continues to increase, silos will continue to encumber technologists ability to meet velocity demands. Developing T-shaped technologists will be an effective strategy to not only develop requisite specialist skills but increase career satisfaction, team cohesion, and speed to market.
Steven Perva is a Senior Mainframe Innovation Engineer at Ensono, a 2022 IBM Champion for IBM Z, and mainframe enthusiast. He spends a disproportionate amount of time learning, teaching, and tinkering with mainframes. He also is the founder of the System Z Enthusiast Discord, a place where mainframe enthusiasts, professionals, and students alike can collaborate, ask questions, and share information about mainframes.
This article serves development purposes only. Practical deployments of this technology should be challenged to adopt best security and integrity practices. Use of any code and commands outlined in this article should be done at the reader’s discretion.
If you enjoyed this article, please follow me for more mainframe (probably z/OS) content here and other cool places:
https://discord.gg/rFXEVKK7AH — Join the System Z Enthusiasts Discord Here
https://linkedin.com/in/stevenperva — Connect with me on LinkedIn
https://twitter.com/SteveMPerva — Follow me on Twitter
Special thanks to Clayton Heathscott and Drew Dunlap.
Seven9s is a proponent of modern mainframe culture. Founded with the objective to advocate on behalf of modern mainframe adoption and the spreading of mainframe culture, Seven9s publishes mixed media, makes merchandise, and is the creator of the System Z Enthusiast Discord.
Get tapped into all things Seven9s here: https://www.seven9s.com
