Introduction
No Country for Old SysAdmins
The System Administrator (SysAdmin) role, part of the Information Technology Operations (ITOps) management philosophy (more definitely defined as part of ITIL), and a long standing and coveted position for many, has quickly and vastly changed in the last 10 years. It is now to the point where it is difficult to even define what a SysAdmin does anymore. While IT experiences evolution from it’s own internal pressures, so too, does it evolve from external pressures, and the average SysAdmin has been under pressure from both sides, such that it is difficult to see a future for the role without new factors coming into play, requiring the presence of SysAdmins within the organisational structure. This is not to say that the SysAdmin is at risk of extinction as an old, archaic way of upper IT infrastructure operations, as recent advances in management techniques, combined with increasingly composable and distributed computing technologies, are slowly bringing back the need for System Administration in some settings, as well as other roles within ITOps.
The System Administrator
What’s the deal with SysAdmins?
Traditionally, while usually focused on the network and administration (e.g. Active Directory) parts of the business’ IT infrastructure, a SysAdmin is a general all-rounder who takes ultimate responsibility for hardware and software assets, and acts as a bridge between the technicians and middle to upper management. The SysAdmin fits into a greater ITOps team within the IT department of an organisation, which provides a service to the rest of the organisation with the IT assets it manages. The ITOps team has three main groups of responsibility; network management, asset management and monitoring/support, of which any or all can be under the responsibility of a SysAdmin, or multiple SysAdmins (Splunk, 2022). The day-to-day responsibilities of the SysAdmin vary wildly, depending on network and hardware architecture, and the software environment from the OS to the programs/services used by the organisation, but generally, the SysAdmin monitors computer systems and oversees software and hardware change. Regarding software, a SysAdmin should have broad knowledge of popular Microsoft Windows services, Active Directory and Group Policy in particular, Unix, enterprise level networking platforms (typically Cisco but dependant on the environment), virtual machine environments, alongside other softwares typically used in the enterprise, like databases. A SysAdmin is also partially responsible for the design of the hardware infrastructure, monitoring systems for faults, overseeing replacement/repair of such systems, and the organisation and roll-out of hardware upgrades or changes. Another significant SysAdmin responsibility is managing data backups and disaster recovery. It is understandable that data loss would at least be undesirable to an organisation, and backups are crucial in ensuring that recovery is even possible. Backup strategies, and data storage strategies in general, have become rather complex in the last couple of decades, and some SysAdmins entire responsibilities are monitoring and managing backups. In larger organisations with more complex environments, there can be many SysAdmins, each with their own subset of responsibilities (Shiff, 2022).
The New Guard
Cloud computing has enabled new ways of fulfilling the service requirements of the organisation, and heralded in the new kid on the block, DevOps. DevOps has come about with the rise of the Agile development philosophy, but is not a well defined thing in and of itself, as, like the concept of a SysAdmin, it is more of a philosophy than a concept. While Agile/DevOps is primarily focused on software development, in the context of infrastructure management, it does away with traditional ITOps roles, as there is only minimal on-premises assets to administer. Instead, DevOps uses quickly configurable cloud infrastructure, offloading the responsibility of ensuring uptime and change management to the Cloud Compute provider. This also allows for the organisation to automate most remaining administration tasks, such as spinning up or down additional compute resources as demand rises and falls (Bravo, 2020).Another similar philosophy, No Operations (NoOps), takes some of the ideas of DevOps to an extreme, by completely eliminating on-premises assets, and delegating all services and support to cloud resources and third parties (Wickramasinghe, 2021). Where DevOps may have had need for a skeleton crew of infrastructure staff, NoOps shifts this responsibility to third party providers.
The CloudTM, but only a little, as a treat
In the last five years, a new buzz word that has come about, “Hybrid Cloud”, looks to moderate the trend of relocating compute to the cloud. Hybrid cloud blends on-premises infrastructure with cloud compute, the benefit of which is lower ongoing cost, better scalability and higher availability (RedHat, What is hybrid cloud? 2018). These benefits are made possible with recent advances in Containerisation, Virtual Machine and automation technologies. Hybrid Cloud can also be combined with another approach, Multi Cloud, which spreads cloud workloads between different cloud providers to optimise for many factors, like cost and reliability. These benefits are possible as some tasks are quite expensive to do in the cloud, particularly storing large amounts of data. Additionally, by offloading some workloads to the on-premises hardware, the number of physical assets (and accompanying administration overhead) is reduced, while using cloud resources to remain scalable to increased demand, foreseen or otherwise.
Is the sun setting on the SysAdmin Empire?
Only a couple of years ago, the future for the SysAdmin, as the role exists today, may have been bleak, but despite cloud revenue continuing to grow, investment in on-premises infrastructure has also increased (Roberto Torres, 2021). This shows that Gartner’s assessment of the cloud vs. on-prem situation does not consider the entire picture, as the Hybrid Cloud model is growing in popularity, and with it, the potential that industry demand for traditional ITOps roles, like the SysAdmin, could at least remain steady. What all of this means for current and future SysAdmins, however, is the scope of the job role is expanding, and is not being broken up into other roles within the DevOps structure. Where the SysAdmin may have less physical infrastructure to manage, the old required experience to manage it is still there, and knowledge in newer technologies will become the new normal, especially cloud environment knowledge/certification. Relatively new knowledge of containerisation technologies, like Docker and Kubernetes is already required in some settings, with programming ability also being highly sought after (Tandog, 2020).
Over the hills and far away
Besides the changes currently happening in the industry, it is difficult to predict how future organisational philosophies and technologies might change in a way that affects SysAdmins. The ideas and technologies already discussed in this essay are already quite mature, even if not widely deployed yet. Where technology and services advance for the enterprise and end user, they ultimately run as software already running in a virtual machine or container that the SysAdmin administers as usual. While containerisation technologies, and Infrastructure as Code (IaC) technologies like Ansible and Terraform have existed for more than a decade, and virtualisation for far longer, these technologies are fantastic for provisioning software onto the available hardware resources, there has been no good way to provision hardware resources. PCI-e Fabric technology has the ability to define a hardware server from a pool of devices, in software, allowing for the automation of reaction of changing workloads workloads on the fly, and allowing for easier physical maintenance of online systems (Liqid Inc, 2020). A relatively new technology that, if becomes more popular within the enterprise in the coming years, has the potential to shift some of the responsibility of the SysAdmin from traditional hardware management to automation tools. The way that hardware is physically managed with PCI-e Fabric would be fundamentally different, too, as individual components are not just hot-swappable in software, but physically too. Data Analytics as a field, is becoming an increasingly important services that companies have demand for, and new ways of crunching “big data” datasets is driving change in IT. AI accelerated workloads are becoming more popular, and, while not immediately important to the SysAdmin as these workloads run like most other softwares on the servers they manage, changes in hardware are coming about that, like PCI-e Fabric, could drive significant change in the industry. Until now, and for a few more years at least, AI workloads have been accelerated by GPU and ASIC hardware, which were already commonplace in the data centre. In-memory compute is a fundamental shift in the way that computation is performed. Where traditional architectures use a Central Processing Unit (CPU), connected over a bus to a pool of volatile memory, like DRAM, instead, data transformations happen in the same place the memory is stored, similar to how neurons in the human brain work. This means that AI neural networks could be accelerated far quicker and more efficiently than on a GPU (Samsung Global Newsroom, 2022). In-memory compute could have a similar effect as PCI-e Fabric to the way hardware is managed by redistributing where compute happens, meaning fundamental changes to the way SysAdmins configure and manage their systems.
Quantum computing, while probably not coming to a data centre near you within the next 10 years, may not be much further away than that, as the technology has been rapidly advancing. A quantum processor is a fundamentally different processor to the silicon-based binary computers in use today, as it uses individual atoms that can perform data transformations in qubits, or in a simplified explanation, two bits. It has already fundamentally challenged the way computing is viewed today, as quantum computing can already break popular forms of encryption. For the SysAdmin, some challenges still exist, as it is not possible to log, in a traditional sense, the state of execution of a quantum state, and quantum storage has not been solved yet either. Speculation over what form quantum computing will take when it becomes a mature, commercially available product is difficult, if it will require entirely new systems or be as simple to install as an add-in card to a pre-existing system, but either way, the SysAdmin need to learn and adapt to the technology (Haller, 2020).
Conclusion
The quickly changing world of IT, fuelled by the needs of the enterprise, forces many to adapt and evolve with the trends of the industry, or become obsolete. Only 5 years ago, it was widely speculated that the role of SysAdmin would largely disappear to automated, cloud based compute, that had no need for a single point of knowledge and responsibility. Advances in SaaS, combined with more complex management philosophies, has seen the emergence of the idea of the Hybrid Cloud, which means that while the tasks and knowledge demanded of the SysAdmin will have to change and evolve with the industry. This means that there will be a definite and continuing demand for a SysAdmin-style role within the enterprise, well into the next 5 to 10 years, and even into the age of the quantum computer.
References
Gartner_Inc. (n.d.). Definition of DevOps – Gartner Information Technology Glossary. Gartner. Retrieved May 9, 2022, from https://www.gartner.com/en/information-technology/glossary/devops
Gartner_Inc. (n.d.). Definition of IT Operations – Gartner Information Technology Glossary. Gartner. Retrieved May 9, 2022, from https://www.gartner.com/en/information-technology/glossary/it-operations
Wickramasinghe, S. (2021, November 29). Itops vs devops VS NoOps: The IT operations evolution. BMC Blogs. Retrieved May 9, 2022, from https://www.bmc.com/blogs/itops-devops-and-noops-oh-my/
Bravo, M. (2020, October 13). From sysadmin to DevOps. Retrieved May 10, 2022, from https://www.redhat.com/sysadmin/sysadmin-devops
Goasduff, L. (2021, November 10). Gartner says cloud will be the centerpiece of new digital experiences. Gartner. Retrieved May 11, 2022, from https://www.gartner.com/en/newsroom/press-releases/2021-11-10-gartner-says-cloud-will-be-the-centerpiece-of-new-digital-experiences
Roberto Torres. (2021, June 29). Spending on servers rebounds as core systems remain on-prem. CIO Dive. Retrieved May 11, 2022, from https://www.ciodive.com/news/on-premise-spend-IT-trends-2021-Forrester/602600/
Tandog, F. (2020, July 31). Sysadmin career trends 2020. Checkmk. Retrieved May 11, 2022, from https://checkmk.com/blog/sysadmin-career-trends-2020
What is hybrid cloud? Red Hat – We make open source technologies for the enterprise. (2018, March 7). Retrieved May 11, 2022, from https://www.redhat.com/en/topics/cloud-computing/what-is-hybrid-cloud
Shiff, L. (2022, March 4). System administrator VS security administrator: What’s the difference? BMC Blogs. Retrieved May 11, 2022, from https://www.bmc.com/blogs/system-administrator-vs-security-administrator-whats-the-difference/
Liqid Inc. (2020, February 19). Independently scale static hardware with Liqid. Liqid Inc. Retrieved May 11, 2022, from https://www.liqid.com/blog/independently-scale-static-hardware-with-liqid
Samsung Global Newsroom. (2022, January 13). Samsung demonstrates the world’s first MRAM based in-memory computing. Samsung Demonstrates the World’s First MRAM Based In-Memory Computing. Retrieved May 12, 2022, from https://news.samsung.com/global/samsung-demonstrates-the-worlds-first-mram-based-in-memory-computing
