DevOps != Products
The last 10 to 20 years in the enterprise has been defined by increasing capital spending on fixed assets (CapEx) while reducing the cost of staffing and operations (OpEx). Gartner report the growth of spend on enterprise application software (EAS) is continuing to grow and in 2015 will grow at 7.5% to $149.9 Billion. According to the Harvard Business Review one in six IT projects had an average cost overrun of 200% and a schedule overrun of 70%.
Some companies who provide solutions to enterprise businesses have continued to use the same delivery model they have used for years, of investing in products (increase CapEx) and reducing people count and replacing high skilled workers with fewer, low cost, lower skilled workers (reducing OpEx). The current levels of under staffing in IT teams lead to constant firefighting that leaves no energy, time or motivation to conduct fire prevention. A common reason heard for buying product to meet a requirement is that the IT staff don’t have the time or skills, but they rarely consider where the skills are created ? Who gave those people the time, scope and resources to develop those skills? Training an upskilling builds employee loyalty, new skills of existing employee allows them to quickly realise business benefit, bringing in new hires with skills which are out of context take longer to realize business benefits.
Other companies have realized the legacy model is broken and have made pivot to lower CapEx spend by the use of open source products hosted on IaaS in the cloud. The effective use of open source products ran to host production services within the cloud increases complexity and as the products are open source and self supported increases the requirement for a highly skilled and highly motivated workforce. The companies who are choosing to invest in people rather than hardware and products to deliver the requirements of business are excelling in the industry as they have improved service agility. Service agility means they can pivot quickly to quickly deliver new features to there customer, they can scale out on demand to give the required density at the right time and also give great visibility. The success of these cloud native companies shows that over-investment in people NOT over-investment in products produces real increases in productivity. The other factor about these companies to note is that with a small workforce of highly skilled engineers everyone is required to cover each other, the developers needed to understand the operational requirements and the operations an intimate knowledge of the requirements of the application. This close relationship and shared empathy between developer and operator is in essence to me what the DevOps culture is.
Choosing people over products is not limited to the cloud native application space, cloud providers make huge investment in hardware and software and the scale at which they work to deliver effectively means they need to draw every ounce of productivity out of there hardware investment. To do this requires again very highly skilled, highly motivated employeers, examples of success in this area have been Facebook and there founding of the Open Compute Project, the Google borg project, AWS and Azure.
The growth of the enterprise application software (EAS) and the purchase promise of the EAS product vendors to reduce the cost of ownership through faster performance, better features, better software and usability and see that the enterprise fails to deliver large projects. Then consider the success of cloud native companies and DevOps culture, it leads us to the conclusion that enterprise IT needs to change fundamentally in how they make decisions.
We can summaries ‘DevOps = People’ and therefore ‘DevOps != Products’