The Good Old Cloud
Has the cloud been here before? Not the actual technology, but many of the ideas behind it. What needs to be done to really leverage it, and gain productivity? What do trends and swings of ideas mean to IT and to IT careers? Will all our know-how be obsolete shortly?
In the beginning, there was time sharing
I started my IT career 1981 in a medium sized, but globally operating Canadian company. We offered access to applications (including E-Mail) and to a programming environment over a private packet-switched network (this network started at about the time when TCP/IP was specified). The end user devices were terminals, at about 6000 CHF per piece.
This type of offering was called time sharing; today it would be Software-as-a-Service (SaaS for the applications) and Platform-as-a-Service (PaaS for the programming environment). It was priced as pay-as-you-go, at 1 CHF per CPU second (on a CPU with less power than that of a modern phone) and 1 CHF per hour of network connection and another 1 CHF per thousand characters transmitted. Flat fees for certain services were introduced later.
Evolution & Trends – Higher Productivity?
The company lost its independence when PCs came up – everybody had their own environment and access to business applications. PCs were expensive, but cheap compared to the mainframe pay-as-you-go prices. With ten PCs, you needed an administrator; with hundreds, an IT department (inverse economies of scale so typical of IT). Networking was in its infancy, with proprietary standards; client-server computing was about to emerge. This brought more complexity, installations, and more IT people.
The individual user, who enjoyed the liberty of the Excel spreadsheets, was often frustrated by the slow corporate IT, which always lagged behind his needs. Indeed, the computing capacity increased by a hundred-thousand, at a tiny fraction of the price. But has productivity increased similarly?
Granted, thanks to the cloud, information will be available anywhere, on any device. But within companies, information is often not in the right form, or confined to silos of applications. Applications processing the data cannot work together without expensive interfaces transforming the data. Each application does wonderful things, but composed together, the productivity gain is smaller than the sum of all individual applications. “Composing” applications and data is not seamless, and the cloud won’t help (especially if different pieces of data reside on different clouds – unless these clouds are linked together in an “Intercloud”).
What needs to be done:
- IT is too much concerned about controlling and baby-sitting the user, rather than about creating value by ensuring processes and data can be composed. Risks and compliance must be addressed, but if users are restricted too much, they will create islands of liberty, making the problem worse.
- Automate everything where human insight does not add value. A lot of “configuration work” falls into this. This frees resources to concentrate on real problems.
- Think about cloud interoperability rather sooner than later. Networking really took off with the Internet standards (IP; Carsten Schloter on IP); the “Intercloud” will prevent many drawbacks of clouds, such as vendor lock-in.
- Apply de-facto data standards that are successful in individual IT solutions (such as RSS, REST, Excel) to the company IT workhorse applications. The integration must be two-way and seamless – the Excel formula must apply within the SAP finance data, not on an extract obtained by a user (who creates a data island).
- Take the cloud beyond the application level: Can I obtain a whole service from a “cloud”; e.g., banking services from a cloud of financial services, rather than from a single bank? Taking this further, how can I compose the services? This opens a new world of imagination.
We’re on a road to achieve what we were dreaming of all the way long – seamless data and computing power for the user:
On this road, there were always swings of trends: From centrally operated mainframes to individual PCs, tablets and phones, and back again to centrally managed cloud services; from control to liberty of the user and back again; from in-house to outsourcing, to off-shoring and back to automation-driven insourcing; from complexity to simplicity, and back again to regain control.
Proponents of newer technologies and trends object this claim, and they’re right – when we swing back, we ought to achieve a higher level, often of possibilities, hopefully sometimes of productivity.
People starting an IT career (and their friends) are often confused by these swings. They are afraid that all their know-how will be obsolete in a very short time frame. This is indeed the case for a lot of specifics; but the main ideas and dreams will stay for a very long time. The principles of what I’ve learned at university and in my first job still hold out, especially now as we embark on the road to this shiny new cloud.