IT Organizational Model Challenges in the Software-Defined World
by Nick Lippis
Cloud-based application development – be it software-defined, containers, or micro-servers – coupled with a new type of organizational model enables IT delivery at the speed of business. Still, one of the biggest challenges in deploying cloud solutions is transitioning to a new IT organizational model, which necessitates a change in existing skill sets and culture.
Traditionally, IT has been organized around technology silos, such as compute, network, storage, applications, DevOps, or virtualization. These siloed IT organizations tend to chain themselves to major vendor products where they receive training and marketable skill sets. But skill sets based solely upon a specific vendor are quickly becoming antiquated. In fact, at ONUG it’s common knowledge that one of the biggest obstacles to deploying an agile infrastructure, which enables business agility, rapid prototyping, new business applications, and downright business competitiveness is this siloed IT organizational model that was put in place some twenty five years ago.
Removing the Road Blocks
Advanced IT organizations understand that during talent and technology transformations, like those occurring now, talent needs to lead. These innovative companies are exploring different IT organizational structures such as a full stack model, in which IT skills are not strictly segmented, but instead are mixed and include a silo skill set along with programming skills. For instance, in this model a network engineer would have some level of overall proficiency in storage, compute, DevOps, or a programming language in addition to his/her deep network engineering skills.
From this data, it’s clear that new multi-disciplined skill sets and tools are needed to design, build, deploy, and manage open hybrid cloud infrastructure. The age of specialization within a single silo is giving way as more and more large corporations adopt open hybrid cloud and private cloud infrastructure. At the same time, the way in which IT service is delivered and managed is fundamentally changing. As a result of these industry transformations, IT organizations prefer engineers with programming skills to those with Command Line Interface (CLI) skills, and a new specialization is emerging.
What’s lacking in today’s IT organization are common methods and programming skills needed to operate and provision hybrid cloud or software-defined infrastructure Cloud infrastructure promises to simplify/abstract IT so that organizations gain operational efficiency as fewer skills are needed to manage more IT service delivery; however, while the industry is on this path, it is not there yet. Cloud infrastructure has not been standardized to the point of commoditization where one person can possess all the skills necessary to do everything and IT organizations don’t need technical staff who specialize in network, security, load balancing, or storage. These skills are still needed. For example, there are people who know how to program or how to deliver infrastructure for trading systems, but they are not the same people who deliver a web front end to the internet. Though they might all use the same C++ tools or open source tools on GitHub, there are different specializations.
This is the change that’s coming to network infrastructure in particular. Understanding network architecture choices is a highly needed specialization within the large enterprise. IT will still need network engineers who know what routing means, and why one would use BGP (Border Gateway Protocol) instead of RIP (Routing Information Protocol), for instance. Still, while we need people who know how routers, switches, load balancers, firewalls, etc. work, knowing their way around a particular vendor’s router or switch is not important.
The ONUG Community loathes to forgo the learning of the last ten years. This history and evolution of network architecture is fundamental to stable IT service delivery. Even to this day, many IT organizations discuss architecture choices, such as the use of spanning tree to deliver IT mobility across a large layer two domain, but forget how this approach does not scale. The ONUG approach to IT organizational transformation looks to build upon this foundation and evolve with changing technologies.
The Full Stack Organization
ONUG’s view on organizational design is based upon the need for multiple specialties and a collaborative culture to deliver cloud-based IT. It’s unrealistic that a true full stack engineer exists; there are too many specialties required for one person. Therefore, this full stack perspective is best manifested in an organizational culture made up of mixed skill sets.
In the near term, many are creating cloud organizations populated with personnel who possess different IT skills in the hope that over time, guided by the right incentives and culture, a full stack organization will emerge. These cloud groups may include those with specialization in DevOps, routing, load balancing, programming, firewalls, and more. Some organizations find that this approach leads to increased specialization crossover. By pooling mixed skill sets into a collaborative cross-specialization culture, in which taking on new skills is encouraged, a new full stack organization emerges.
Another approach to integrating specialties has been for each silo to donate one of its personnel to the new full stack group, which starts to build upon different knowledge bases. This knowledge base expansion becomes the root of the culture and fundamental to the full stack organization. The first phase of this transition is obtaining agreement upon donating the specialists with the overall goal being that the knowledge base will expand significantly and permanently over the next five years. The full stack organization is focused on collective skill sets and the creation of a deep knowledge base that’s equipped to design, build, manage, and troubleshoot cloud based applications that span the full stack.
While it’s not the expressed goal of a full stack organization, there is an expectation that as cloud-based tools become available and the knowledge base becomes operationalized, IT organizations may not need as many specialists. Therefore, the number of specialized IT operational engineers needed to safely manage and support the same amount of infrastructure devices will decrease. For example, at ONUG today, one network engineer usually manages less than 200 devices, but in the future that engineer should be able to manage thousands to tens of thousands of physical and virtual devices, and those devices could be network, storage, or compute.