In technical world, any paradigm shift has stepped up the technology evolution to its core. The transition into a new paradigm is interesting to watch, specially when people will continue to follow their know-how proven-correct approaches while wanting to ride the new realm of possibilities the actual shift brings. Doh!
Paradigm shift by definition is “A fundamental change in approach or underlying assumptions and leading throughout a paradigm shift with the ones' learned old assumptions not only hinders the progress but also is the central reason for waste of time and vital resources.
For example if you consider notion of Cloud computing, which was not a single solution but a paradigm shift in how one can achieve (perceived sense of unlimited) scalability, flexibility and on-demand usage of resources even though resources are in fact limited. There were many tools and components (including virtualisation, as a service business model and more) that made this paradigm shift a liveable reality.
Every industry started to think about how to adopt to this new reality, while some captured the true essence of it, others still continue to think as they did prior to this shift.
Cloud in other Domains
In many other domains the scalability, flexibility, and on demand use of resources is achieved already. In telecommunication systems, the notion of scalability, flexibility and on demand allocation of resources was put in place for radio resources long time ago even though it was not called Cloud. This was possible by various concept of time, frequency or code division for multiple access to radio resources. This flexibility and scalability is well embedded in the network, and the operating system of the network makes this on demand handling of network resources possible similar to the way a hypervisor will provide isolation of virtual resources in a logical server.
In non-technical systems, for example in organisational leadership, the notion of scalability, flexibility and on demand allocation of (human) resources, is put in place in a very primitive way by allowing the build up of "virtual teams". The virtual teams are supposed to come together on demand and form around specific tasks. In an organisation, the operating system would be the management system. It has been always fascinating for me to compare the human driven systems with technology driven systems. Even though the technologies are build by human, the human driven systems most of the time are less efficient than technology driven ones.
Scalable and Flexible Operating Systems
The isolation and on demand allocations of radio channels in the network infrastructure is done with the help of a well defined, fixed and robust algorithms ( inside operating system ) in the network infrastructure. The "response time" of these computations has to measure up to the precision of requirement on radio channel allocation and therefore cannot be a random parameter.
Similarly the computing in a server operating system or a hypervisor or a cloud management system such as OpenStack is pretty fixed and robust in order to manage the on demand processing and allocation of resources in a server. The question is whether it is worthwhile to consider the scalability, flexibility and on demand concepts for computation at operating system level, regardless of whether it is the operating system in the network or a server ( or a management system in an organisation ! ).
On Demand Management System
Answering the question of scalable operating system in the network or a server requires a deeper look at the operating system functionalities to see which part really benefit from scalability. In turn that would require more modular build up of an operating system. In today various operating system, there is a high level division between user plane and control plane and control plane is normally the fixed part and user plane is subject to flexibility. I leave that for another time or to your own curiosity to look into operating systems and see the bottlenecks in areas that would have benefited from scalability.
Now, just for fun, consider an organisation and its management system. Across all industries, there is still high attention on management system which are pretty fixed. Imagine if a management system also was seen in the light of control plane and user plane. Then control plane, would be taking care of all control mechanism ( financial, legal, and so on ) and the user plane was dealing with all the operational tasks. It is normally the operational part of any management system that suffers from lack of scalability and could be interesting to experiment with an scalable flexible on-demand deployment.
New Paradigm, New Assumptions
Regardless of the system you are dealing with, it is worthwhile to examine the current assumptions and approaches, and before turning on the switch for any new paradigm, check the assumptions and approaches which you have to adapt to as part of the journey.