Software development is not a new profession, people had programmed systems – analog and digital – for a time much longer you will expect. I think the earliest program I know is a time measurement program, developed by some ancient greek and realized in an analog way by letting water drip out of reservoir with an individual diameter as input and having a meter as a display. But let us focus on the electronical side where we have started with electronic tubes, going to transistor based host systems and further via workstations and PCs to client server based systems and finally the cloud as a service provider. This is huge development, but what cuses it? From my perspective it is not only the availability of technology – you could still use a dumb VT100 Terminal to book a holiday home on a host computer named airbnb. And it is not only the availability of a solution – you could also connect a terminal via phone line to the airobnb server much cheaper than using your iPhone 12 with you expensive flatrate. I do believe that the driving force for changing operational models of software is the efficiency of their development.
In the beginning there was the host. He was hard to program because he could do only one thing at a time. He had only one user. And every change in a program – even a tiny change in a subroutine needs a shutdown of the whole software and sometimes the whole computer. Not very efficient.
Then computers become smaller and more easily available – maybe you remember the MicroVax or the SGI Indigo. But more people got access to this compute power, so the demand grew. Because of this the use of computers rose, but software development became less efficient. For each change in a software you need to distribute the software to millions of computers and force the users to reinstall the software. That is when patch days were invented. And people hated it. In retrospective view it was a stepback from the host computer. This is why it is mostly gone
In the same time the network was invented. The first use was the transfer of information – mail and ftp. And telnet brought back the host and the efficiency of host based software development. And client-server based computing tried to make host computing more efficient. This was done by creating fat clients, which took all the standard functions of a software which is changed seldom to the workstation and leaving the core functions, which are continuously developed, on a host-like server. This eases development a lot because changes could be implemented much faster.
In the meantime the network developed and many people started to operate their own hosts – accessing them with the actual version of a dumb terminal, the webbrowser. And people found out that they can benefit from other people work by linking alien services in their own solutions. If I access my favorite news aggregator I use the ID of my favorite search engine as a login so the developer of my news aggregator does not need to take care of developing and operation the ID management which makes his development much more effective. I call this cloud usage and it differs significant from just using somebody else’s datacenter.
This is the reason why I think that the move to the cloud is indispensable. Not because people want it, but because the development of solutions is more effective, therefore faster and therefore brings more value earlier to the user. Ignoring this leaves you behind the digitization front. And therefore the transformation into the cloud is indispensable.
Der Beitrag erschien zuerst auf linkedin.