Let me start by outlining what I mean by "cost of local computing."
Of course, we have to purchase powerful devices to execute local computing – be they laptops, desktops, servers, mainframes etc. Broadly, I'm thinking of business uses here, but home / personal computing will follow a similar path.
In addition to specific purchase costs (and the capex implications that go along with that), we have to deal with a huge variety of support costs – network infrastructure, firewalls, local security (including, obviously, anti-malware scanning and its ilk), etc.
Everyone has to buy and manage their own back-up environments and everything else required to maintain good order.
The big problem here is that the purchase and support costs are high, and the dangers equally high. The IT press is filled with stories of "bad dudes" doing all manner of nasty stuff. Everything from complex hacking, all the way down to ransomware and crypto-jacking.
Take ransomware for example. Almost invariably it is the random low-level employee who opens an attached file, which, rather than being the finance sheet they were promised, will launch the attack and soon enough, much of the organisation's data has been encrypted.
Of course, we blame the employee for not being more vigilant, but they should not be regarded as our primary line of defence – it's NOT their job. Most anti-nasty software seems incapable of detecting or trapping this rapidly evolving family of malware. Sure there are a couple of exceptions – Carbon Black and Cylance spring to mind but as I see it, they're still attacking symptoms.
What if just about all of that went away?
What if the device on your desk was something akin to a Chromebook? (I use this as an example, not as a specific recommendation). This device has no local storage and acts simply as a remote terminal to a Cloud-based server – in effect, some form of remote access to a virtualised environment running somewhere on the Internet.
This gives us a lot of advantages. A combination of automated user support, software-defined networking and highly optimised load management will mean that most of the corporate IT support costs vanish. Remember that in all likelihood, the hugely powerful CPU in the device you're reading this article on is barely averaging 1% utilisation over a 24-hour period
Further, it is a lot easier to deploy universal anti-malware in a co-ordinated fashion when all email and other access is centralised and machine learning systems have access to the entire spectrum of nasties.
Data storage is guaranteed to be safe and fully protected with millisecond scale back-ups available to all users. User security can be more tightly managed from a centralised view.
What do we need to make this happen? Mostly, just "political will" from the senior executives in any organisation – the desire to pass management of their computing infrastructure to someone else. Most will argue that they're not in the IT business, yet will argue in parallel that IT is a core function.
Technically, most current Internet connections will be sufficient to support typical usage – documents, spreadsheets, database work etc. Probably the only users requiring something close to gigabit connections will be image processing workers (PhotoShop et al) and gamers.
Think about a couple of scenarios. As a sales executive, you walk into a client's building, sit your device on the presentation lectern; it automatically connects to your remote workload and you're ready to present your slides in seconds. In this instance, your device may actually be your phone.
Or perhaps, as a contractor, you work with a number of concurrent teams, all on unrelated projects. By executing a context switch on your device, you can swap between environments almost at will.
When could this happen? Actually, it already is.
Many people are making use of cloud storage to ensure they're always connected to their data, no matter where or how they access it. The proliferation of XaaS regimes means that we are already familiar with the concept. Further, organisations such as AWS, Google, Microsoft, Facebook and SalesForce (I'm sure there are lots of others) are already "training" us to accept this new remote-computing paradigm.
As the danger of malware increases, and the ability to defend against it decreases, this solution will become more and more attractive. I'd estimate that within 10 years, the vast majority of workers will be cloud-based, working in offices with zero computing resources, or IT support teams.
I'm very keen to hear feedback on this – please respond in the comments below.