IT's vicious cycle
Because of current trends such as digitalization and mobility, existing companies need to constantly change their business models. There are many examples where a company went out of business because of not doing this. Positive examples such as IBM or Nokia proof that it is possible to change even big corporations. They succeeded in reinventing their business model according to the changing context and found new revenue streams keeping them alive.
This ongoing and accelerating change of business models has dramatic consequences on how to build IT systems. Todays legacy IT systems are mostly huge monoliths. It is hard to change parts of such a system, because of many intra-system dependencies and big test efforts. Therefore, not many companies are able to deliver new releases of their software solutions within a couple of days. But the business requires that the IT systems need to be adapted quickly and continuously.
This situation leads to a vicious cycle, where the IT organisation gets slower and slower. There is not enough time to do it right, shortcuts are taken, and technical debt is accumulated. Eventually, this leads to unstable and non-performant systems.
The lack of skilled IT resources aggravates this situation even more. To build a distributed system in a service architecture needs different skills than to build a monolithic desktop application. The progress of technology in this area is astounding, but there are not enough technical people to apply the new technical solutions in projects. Some try to solve this with outsourcing and near- or offshoring. But this is a short-term solution at best. The external IT resources are getting more expensive when needed, core knowledge is lost, and the projects get more complex because of different cultures, languages and time-zones.
It is common knowledge that IT systems should be built in a (micro-)service architecture. Such an architecture decouples parts of the system using networked APIs. This enables to change parts of the whole system with minimal impact on the rest. As long as the interface of the service stays the same, no changes are needed somewhere else and the new implementation can be locally tested. Startups proof that this is the right approach, but they have the benefit to start in a green-field approach.
For established companies, it is difficult, costly and time-intensive to renovate a legacy system into a new architecture. To build a new system from scratch besides the currently operating one, binds too many scarce resources and is high-risk because of the high costs and the project risks involved. Therefore, many IT organisations are stuck in legacy.
When the business is not satisfied with its own IT (e.g. because of high cost, low efficiency, poor quality, or slow speed to market), it selects solutions where results seem easier to achieve, especially cloud solutions. But uncontrolled usage of cloud services today leads to even more complexity and less security. This is because such services are not integrated with the rest of the enterprise IT. Many passwords need to be administered and data is stored externally in several places. Because of lacking IT knowledge, business chooses risky systems that may lead to problems later. To use cloud services in such an uncontrolled way leads to chaos. Such a situation is called 'Shadow IT', because no-one really knows which services are used, where the passwords are or where the data is stored. At first sight, cloud services seem like a promising idea to deliver better time-to-market, but over time one realises that they lead to growing costs and less security.