BankerVision

An update on my alternative operating system

A few weeks ago I posted that I was getting rid of Windows, and I thought you might all like an update on where I am with that.

The biggest change for me is that I've started to run Linux under the hood on my principle work laptop. I'm using Suse 11, for those who are interested in technical things.

Originally, years ago, I had a Macintosh. It was the original 128k version, with a single floppy disk, and I loved it. I continued with Macintosh for the next decade or so, way before it had its resurgence to popularity in recent times.

I gave it up and switched to Windows in order to be more aligned with what most corporates were using at the time. Over time I became pretty convinced that Windows was the future, and even wound up working at Microsoft eventually.

I've now realised that Windows isn't cutting it very well any more for me. The realisation came when I couldn't get the thing to go to sleep when I wanted it to. Neither could I get it undock without a restart. Couldn't connect to any of the printers at work. Couldn't even reliably sustain a connection to my printers at home. Had to beg it to make a connection to a wireless network. And the integrated Internet Explorer was crashing at the drop of a hat.

I realised I was spending a lot of time fighting with it before I even got any of my applications to open.

Yes, I know these are all signs that it's time to reinstall. But stop for a moment. When did we all get into the habit of reinstalling operating systems because they started to degrade?

It was the high water mark of inconvenience that made me jump ship.

Of course, I still run Windows for those key applications which I can't do without. For example, I can't live without ClearContext, which helps me manage email. And I still need Windows to access the corporate resources at the bank.

So I run it in a Virtual Machine on top of Linux.

Here's the surprising thing. Windows runs better inside a Virtual Machine than without it. Even from a performance perspective (although my laptop is a relatively new one and has that new Intel virtualisation gadget in its processor, so maybe that is the reason).

Even my Microsoft Office applications seem to crash less frequently. I still use them, having tried and failed to convert my brain the OpenOffice things that came with my new operating system.

I now have a copy of Windows which I run every single day, and I haven't had to restart it in over two weeks. Unheard of. Neither is it degrading with time. Bliss.

And guess what?

When you undock the laptop, it undocks. When you ask to sleep, it does so in seconds. And comes back to life without hassles.

Now let me not paint a picture that all this was simple to achieve in the first place. I had quite a lot of bother getting everything the way I like it and went through two Linux distributions before finding Suse 11.

But, and I have to say this, I'd evaluate the bother as about the same as making Windows work perfectly, all the time. Especially when you factor in the several times a year reinstall.

For me, the switch to Linux is paying off. I am actually more productive (at least now it is all working) than I ever was when I had to fight with Windows every day.

Counting up the minutes wasted by operating systems is an illustrative exercise for everyone. Keep a wasted minutes diary for a week. You'll be shocked at what you find out.

Self organising and optimising business process

I've spoken here often of the democratisation of the tools of production, something that is happening across so many technological and business domains. We are entering an age where anyone with the will to do something has the tools to do so. Talent becomes the key differentiator rather than access to capital and resources.

Anyway, yesterday I was talking with a group of people here at the bank about this kind of future, and we got into a discussion about what happens when we put the tools needed to created business processes into the hands of anyone at all, something we are just starting to do here, by the way.

In a command and control organisation, of course, such a concept is intolerable. For every given process, there must be one true way of doing things. Because it is the one true way, every single eventuality has to be understood up front, and a rigorous change process applied to make sure things don't break.

But when anyone can create a process, you get proliferation of different ways to do the same thing. Holiday leave applications and tracking? Everyone can build a way of doing it, and continue to use it whilst ever it is convenient to do so.

The thing is, business processes of this kind can be considered to be social media as much as a blog or wiki page. They attract audiences (users) and the ones with the greatest numbers are, ipso facto, the ones that an institution uses to move forwards. By attrition from lack of use, bad processes go away by themselves.

In other words, the set of business processes that make up an institution comprise a self organising, evolutionary system. In nature, such systems respond very flexibility to changes in their environments, moving immediately to the most optimised use of resources available.

Let there be multiple ways of doing the same thing, and let the end-users decided with their feet which ones they like to use! Its how open-source has worked for years now.

That's quite a contrast to a monolithic process established on behalf of users by some faceless re-engineering entity, with little or no consideration of whether the process is actually convenient and works. And God forbid if something changes, because then its necessary to get a budget, start a project, and engage a (probably) overworked IT team.

Now I am not suggesting that major customer facing business processes should be self-evolving and user generated, at least not yet.  But the fact of the matter is that around the edges of any major business processes, there are already lots of different ways of doing things, all dreamed up by workers to make their jobs easier.

How does a paper based customer record get from the fax machine to the computer system in the branch? Probably via a folder marked “for entry” in some branches and done in batch, and carried immediate it arrives to a terminal for keying in others. Individuals and work groups create processes for themselves all the time which are optimised for the work, and styles of working, they prefer.

The democratisation of the tools of production of business processes adds to what is already going on in one simple way: it makes it possible for any worker to publish their way of working to others.

That's a fundamental new ability, because it means that the best workers can replicate what makes them successful. Its a capability that will give any organisation who cracks the code of doing it a significant competitive advantage.

Where is my traded market for cycles?

There's quite a lot activity in the cloud space these days, what with Amazon, Google, and a pile of others entering the market. And it is probably quite true to say a  few institutions are considering how they might use these services in substantive way down the track.

Let me make my own thoughts on this clear. In 10 years, 2 decades at the most, I believe banks will own neither data centres or networks. Networks are expensive things, and you always run out of money before you can provide enough bandwidth to everyone who wants to use them. At least, that's what happens when you build your own ones. Public networks are quite a different story, though.

And today's data centres, fixed to specific location, are almost never efficient to run. But new ones are springing up with access to environments useful for cooling, and lovely green hydroelectric power.

Let us now imagine the imposition of a national carbon emission tax on IT, which some countries are already considering. The impact is that inefficient data centres would get expensive very, very quickly. Why on earth would you own one with all that kind of uncertainty about it's future?

And a bank of any scale owns several data centres, not just one.

In any event, it will be economics of running infrastructures in a climate of energy price instability and regulatory uncertainty about carbon that force institutions to use cloud resources for core processing. I'd not be surprised to see banks putting their processing in locations that optimise the cash return per cycle, and switching locations around rather often.

You will be able to do a virtual data centre soon, one with all the security and enterprise operations features that banks presently demand from their private infrastructures.

But the point is that once you have a virtual data centre, able to move its physical location around as needed, you then need some certainty about its future costs of operation. Data centre operators are subject to significant variability in their cost of operation: their key input is energy, which is price-volatile (and becoming increasingly so).

So, where is the emerging traded exchange for cycles? Where are the futures contracts we'll want to get a price optimised across the decades we have to run systems? Are data centre operators even thinking about the financial instruments we'll need to have in order to use their shiny new facilities for anything at all mission critical?

I don't think so. At least, I haven't seen much evidence of it. But one thing I can say: bankers I've spoken to in lots places around the world are wondering at their absence. Perhaps it is an opportunity we -as in financial institutions collectively - pursue independently of the data centre providers?