Transforming Disciplined IT Functions: Guidelines for DevOps Integration

Introduction

In response to the many opportunities and challenges of this digital era, many large companies have been rethinking the organizational setups of their IT functions. Digital innovators such as Google and Amazon regularly release information about the newest trends in technology and serve as role models for other companies. Furthermore, large organizations manage their departments as organizational silos in which development and operations employees tend to be highly specialized. As a result of the silo setup, strict managerial hierarchies and clearly defined work environments often make communication across IT subunits rare. A key task of IT departments is to provide new software while ensuring smooth operations and functionality. In the silo setup, software updates are typically planned, constructed and executed as waterfall-style command-and-control projects (Dery, Sebastian, & van der Meulen, 2017), new software functionalities are seldom provided and large-scale error-prone releases are often executed on the weekend to avoid extensive system outages during peak business times.

As companies adopt enhanced development processes to make projects successful, it has become clear that providing complete service lifecycle support is most efficient when operations and development activities are combined. Developing capabilities and collaboration across internal IT subunits is necessary for both clear intra-organizational communication and improved performance (Dhaliwal, Onita, Poston, & Zhang, 2011). As organizations realize that silo IT functions are less able to leverage digitalization to maximize business value, they are more frequently integrating cross-functional IT teams in their IT departments to ensure fast and stable IT services (Wiedemann, 2017). The concept of DevOps is widely used to structure cross-functional teams with end-to-end responsibility for one or more IT services (Debois, 2011). In fact, a 2017 global conducted DevOps benchmark study by the market research institute Forrester presents that more than 50% of the participating organizations have implemented DevOps by end of 2017 (Stroud, 2017).

DevOps combines agility and rigor in cross-functional teams. Agility is defined as a method that enables rapid reaction to changing user requirement in an iterative way (Hemon, Monnier-Senicourt, & Rowe, 2018). The architecture of established IT departments often yields monolithic IT system with tightly coupled centralized application systems and bulky, often rigid legacy systems. Although this structure worked very well for a long time, progressive organizations following DevOps principles are striving for greater agility. For example, by creating internal IT ecosystems with loosely coupled IT infrastructures like microservices for managing frontend systems that run on digital service platforms serving as backend systems (Ross et al., 2016). Process rigor is defined as the stringent enforcement of and adherence to compliance and formal, structured IS processes. Process rigor builds upon formal, clearly defined, and detailed explanations of tasks, roles, activities, work products, measures, and methods related to IS software processes. An example of a plan-driven software development approach based on process rigor is the Capability Maturity Model Integrated (CMMI) (Lee, DeLone, & Espinosa, 2010). One way to achieve process rigor in software operation is by implementing the highly formalized Information Technology Infrastructure Library (ITIL)¹.