Only a few companies have had good experiences with the introduction of DevOps in traditional development and operational units. The goals of the traditional Dev and Ops teams are simply too different for that. However, the multi-cloud can provide new momentum. Multi-cloud and DevOps are like two sides of the same coin.
Multi-cloud architectures have also arrived in the German SME sector: according to Crisp Research, 85 per cent of small and medium-sized companies in this country are already working intensively with the cloud. And almost two-thirds (65 per cent) are using hybrid or multi-cloud. These companies are thereby creating the best conditions for a DevOps culture, in which development and operations work closely together instead of being strictly separated from each other as before.
The new momentum from the multi-cloud is also urgently needed. Experience shows that DevOps initiatives within existing departmental silos are often doomed to failure. While the classic task of developers is to achieve innovation, operations has so far been primarily concerned with keeping the systems running. These sometimes mutually exclusive goals extend to employment contracts and cannot simply be made obsolete by introducing new rules.
When errors are suddenly welcome
DevOps also involves a fundamentally different work culture – for example, when it comes to the topic of ‘error culture’. Traditionally, entire process chains, team rules and even remuneration issues are regulated according to the principle of the ‘blame game’: first, responsibility for errors must be clarified; the solution comes second. DevOps culture is diametrically opposed to this approach. In DevOps culture, a particular error may be very welcome because it advances the entire process in the sense of an exploratory mode that follows the principle of trial and error.
By contrast, the multi-cloud virtually demands overarching collaboration and can thus reduce the potential for conflict between development and operations. Containers are the best example of this: companies get the most out of their IT infrastructure when they make themselves independent of providers and operate a multi-cloud structure with overarching container clusters. This in turn reduces the friction between development and operations. Developers have complete freedom within containers and can still be sure that their applications or microservices will work wherever there are suitable interfaces. This significantly increases the quality of deployments.
A joint approach to multi-cloud and DevOps
Notify me of any new publications
Conversely, anyone who uses multiple cloud offerings and continues to rely on traditional development methods such as the waterfall model is acting like someone who buys a Porsche and then installs the engine of a small car. In the words of the US developer magazine Infoworld.com: ‘The cloud is actually like an engine. If you don’t use DevOps as well, at least one cylinder won’t fire, and probably several. And if you use DevOps without the cloud, the result will be just as poor. Cloud and DevOps have a symbiotic relationship.”
These changes in underlying conditions represent a particular challenge for medium-sized companies. That’s because their software development is generally based on needs that differ substantially from those of international groups. If companies nevertheless wish to benefit from the prices and added value of best-of-breed technologies, then there is no alternative to a multi-cloud, for which QSC, for example, offers operating, management and advisory services. QSC then not only provides medium-sized companies with
DevOps metrics drive change
Alongside multi-cloud, key figures and metrics are ideal drivers for DevOps initiatives. They are indispensable for ensuring transparency regarding the current state of affairs. That’s because DevOps can be measured – for example, the frequency and methods of communication: Do Dev and Ops hold regular meetings or do they only exchange sporadic e-mails?
The CALMS framework for DevOps is a useful tool for selecting and defining the most important metrics (CALMS: Culture, Automation, Lean, Measurement, Sharing). It defines the characteristics of DevOps that are crucial for success. On the basis of the defined and collected key figures, measures can then be derived to drive the desired progress. What is important here is that all the measured levers cannot be ‘worked through’ one after the other, but must be tackled simultaneously, as they are interlinked and mutually dependent.
But it is precisely those medium-sized companies that have been established for many years that are finding it particularly difficult to introduce DevOps. This is because employees there are all the more likely to stick to familiar paths and act according to old patterns.
