What three billion-euro deals in open source mean for systems engineering
There is no doubt that open source is largely responsible for the fact that we are experiencing an unprecedented variety of new and innovative products. This is because open source supports unbureaucratic reuse and collaboration across organizational boundaries.
Open source has also featured a lot in the business news this year: This year, after all, there were three billion-dollar deals involving open source, and three acquisitions at that: IBM bought Red Hat ($34 billion), Microsoft bought gitHub ($7.5 billion) and SalesForce bought MuleSoft ($6.5 billion).
It's obvious: open source is big business. But what are the implications for systems engineering?
A stable infrastructure
Software is one of the reasons why systems of today's complexity can be developed at all. Every variable in software represents a "moving part" without us having to worry about wear and tear or failure. A mechanical watch consists of at most a few thousand parts (not all of them movable), while a modern car hundreds of millions of lines of code can contain. (Now it would still have to be estimated how many variables, i.e. moving parts, are in it - but the point should be clear.
The software architecture extends far beyond the product software. A simple smartphone app itself generally uses many software libraries that set the framework for the application (most of them open source). The app is written in a standardized programming language, for which there are certainly open source compilers that are often used. The whole thing runs on an operating system that is either open source (Android) or processes many open source libraries itself (iOS). In order to function, an Internet connection is required, which is operated via a whole army of open source tools, such as web servers and DNS servers, which in turn run on servers with open operating systems, often from Red Hat.
Systems of Systems
The infrastructure just described is also an impressive example of a "system of systems", i.e. loosely coupled systems that display new properties when they interact. Open source is ideal for this, as the interfaces are visible and accessible to everyone. And if interfaces are missing, they can be added unbureaucratically.
As many of the market participants are aware of this dynamic, there is now a consensus that open source requires governance, which manifests itself in the form of foundations (such as the Eclipse Foundation). The exciting thing here is that the corporations are performing a balancing act: On the one hand, they would like to have control over the software; on the other hand, they need acceptance in the market, which is held back by the market power of the corporation. For this reason, corporations often become active themselves and voluntarily relinquish control. This is how the Eclipse Foundation was founded by IBM.
Open source is socially acceptable
Companies are now using open source without hesitation, it has become socially acceptableThe added value is clear to decision-makers. For systems engineering, this means that the use of open source is strategically important: the potential savings compared to proprietary solutions are enormous, the disadvantages minor. However, aspects that relate to differentiating features or are part of a company's intellectual property should only be placed in open source with caution, if at all. But the majority of developments are generic, and the added value of open source is beyond question.
How not to do it
As a cautionary example a German bank banned its suppliers from using open source. Such a requirement shows a gross lack of understanding of modern software development. Apart from the fact that this drives up costs for no good reason, the software developed in this way is certainly also less safeis an important criterion for a bank. I can only agree with the blog author, Dirk Riehle: Anyone holding shares in this bank should sell them quickly.
