moss.geek.nz

I attended CloudCamp Auckland today and came away with a bunch of notes and ideas and concepts which were different from what I had going in about this nebulous realm of “cloud computing”.

Let’s begin talking about it by defining briefly what we mean by “cloud computing”. The cloud is the part of diagrams you see in computing that’s simply “technology that’s not your problem”. You know, the part that says “Internet”, or “Network” or “Interface” .. between the lines that link the little computer icon or laptop icon with the circular data storage icon.

Cloud computing is best understood as a layer of Information Technology. We all have some experience with different layers of computing – 386s with Windows 3.0 versus Pentiums with Windows 98 versus Core 2 Duos with Windows 7. Dialup versus Broadband. As we move up the layers or levels of technology, we are enabled to do more with what we have. By the same token, you can also travel backwards and for the most part do all that is enabled the old fashioned way.

The main trend in computing in the past twenty years has been digital wastage. By digital wastage I mean that we have more and more computing power, but we always seem to be doing less and less with it. As I type this in OpenOffice Writer, my computer is busy deciding cycles to mundane functions like pretty graphics, checking for anything that may interrupt me, and so forth. I could just as easily be writing on a mechanical typewriter – but the word processor enables me to do more.

A simple way to think of the wastage is in the levels I presented earlier – pairing a version of Windows with a processor. We haven’t seen Moore’s Law – that the speed of computing doubles every two years – be particularly relevant in the realm of consumer computing. We find our old computers, which were cutting edge, be described as boat anchors; while they still function roughly the same (ignoring that consumer computers generally become slower over time as more and more programs are loaded on them that further waste resources). It seems that computer processing fills up the speed that the processors provide.

And what cloud computing does, is takes that and makes it not your problem. You don’t need to put up a huge capital investment and get a mainframe, or get a high end box on every desk, or build a server room, or even your own data center – you can go and pay a cloud computing supplier for just the computing you need.

This is the value and advantage of cloud computing. Think of it like a cellphone plan – there’s a bunch of different big providers out there, and they give you a range of plans to suit your communication needs. Except, this is for computers.

Let’s take a simple example. You have a machine that can handle 200 people. If you provide your own computing, you’ve had to buy that machine, put it somewhere, and maintain it’s connection out to the internet. It’s a lot of “sunk” costs. Now, 201 people show up. Oh dear, you have to go out, buy a new machine, stick it next to the other one and make sure that it’s got all it’s maintenance too.

Or, you could be using cloud computing. You have a machine. Somewhere. It’s not your problem. It serves 200 people. 201 people show up. You’ve told your cloud computing people that when that happens, please give me another machine for a while until we’re down below 200 people again. And they can provide you the new machine within minutes, and for a specific length of time.

That’s the three things that really make up cloud computing. It’s cheaper and easier than doing it yourself. It’s just the computing you need. And you can get more resources within minutes.

So now you ask, well that sounds great? But what’s the down side?

Well, the down side is you don’t own the machine. It’s somewhere else. And you don’t know exactly where that is. In most cases, it’s not even in New Zealand. Which opens the two big can of worms;

Number one, is my data safe?

Number two, can I rely on the cloud?

The answers to both of the above questions summed up in one word is Yes. But that’s summing a lot of very real concerns into basically a faith about the underlying technology.

Yes, we can make your data safe. For incredibly paranoid levels of cryptography. But if someone’s password gets stolen, you can’t just lock the doors of your building and make sure no one suspicious is sitting down at your computers. And no, your data is not physically in one location. It may even be overseas and subject to various international laws.

Yes, you can rely on the cloud. For very high levels of reliability. But if it goes wrong, it’s not your problem, so you’re not part of solving that. If your cloud based service goes down, all you get to do is pick up the phone and yell at whoever provides it for you – and once again they may be on the other side of the world from you. And there’s usually less people managing a cloud for more businesses than if you had your own gear and your own guys.

But in saying that, cloud computing is by and large very reliable. By the very nature of being vast and scalable, cloud computing providers have a lot of infrastructure. And they exploit economies of scale on that infrastructure. Reliability is essentially a technical issue that with a cloud, becomes somebody else’s problem, and that somebody else has the resources to solve it because they’re dealing with a lot of computing power.

So, we really need to focus down on the safety of your data. Essentially, your data safety is still your problem. You can’t delegate that to someone else. So you’ve got to think about it and deal to it in some way, be that with good technology or good practices.

And we’ve got to get over the idea of physical control. Physical access to data means nothing nowadays. You may physically control your computer, but it may be infested with viruses, or have unpatched vulnerabilities. Just because you can point to a physical drive with data on it, doesn’t mean it’s any safer than being in a cloud. It’s really up to the technical community to get this idea across to everyone – where your data physically is, isn’t who controls it. You control it. And if you’re not happy with what safeguards a particular implementation of cloud computing has over data control, then vote with your patronage and go somewhere else. Or ask the provider or for technical help to control it to your liking.

The one aspect of physical control that is still relevant is regionality and nationality. Especially to us here in New Zealand. There’s no way of overcoming our geographical distance from the rest of the world. And in the digital age, it still translates into a latency issue – no matter how fast the connection to America, it can’t go faster than the speed of light across the Pacific Ocean. And we’re not near that yet, but there’s definitely a direct link between distance and time in the digital realm over long distances.

So the question must be asked, Does New Zealand have, or even need a local cloud?

Well, first off, we don’t have a local cloud. We don’t have the infrastructure. We’re showing ourselves as being quite adept at using cloud infrastructure, but we’re generally using American service providers. And there’s some work on cloud providers in Australia that we’re working with as well. But there’s nothing locally.

But do we need it? Well, that’s a subject of some heated discussion. The answer among cloud computing enthusiasts is, Yes, we need it, but No, it’s not viable to build it.

Why do we need it? Because New Zealand is positioned globally in a precarious spot. We’re an export nation that is getting outbid and outmaneuvered in the global economy by the sheer scale of everyone around us. We’re only four million people. 98% of our business is small business. And by small, we mean they don’t even register on the scale overseas.

And the global economy of scales destroys the idea of a local cloud as well. It’s estimated that it costs around three times as much as offshore to provide computing resources in this country. Because, well, you’ve got to import the gear, you’ve got to set it up, and then you’ve got to pay to get the people to stay here to work on it.

But we need a local cloud because of our position amongst the world. We’re small, agile and lean. And that’s where cloud computing really comes into it’s own – it provides what used to be enterprise level computing on demand. So small, agile and lean companies can play with the big boys. And we can extend that to countries.

Which is why we should have a local cloud. But not just any cloud – we need a testbed cloud for development. We’re an export nation at heart, and we don’t have the local market to support the sort of scale you can push something to through cloud computing. So what we should have is a way to quickly develop and test products here, locally, and then push them overseas and scale them up there.

Because there’s no point in rerouting American customers from American clouds to ones on the other side of the Pacific. And no one wants to send their European customers to a cloud hosted the other side of the world away from them. So we should put our work out there in their clouds. But, we should work on it here in the Land of the Long White Cloud.

And there’s another great advantage to this too – we could work out the kinks of pushing something from one cloud platform to another. Because there’s a lot of different technical platforms out there – and one day, the need to bring them together and migrate from one to the other is going to be huge. So why not solve it, and then sell that solution?

So we need, and we can build a staging cloud. But who’s got the resources to make it happen? In New Zealand, we’re really lacking in our incubation sector. There’s not much Venture Capital around. Who’s going to sink the cost of making a cloud? Well, it’s not our big companies. They’re not IT shops. It’s not our government’s job. Or is it?

Well, here’s a proposal. We’ve got this idea of a state-owned enterprise. Why not a state-owned computing cloud platform? That is, if we could get the legislatorous to wake up and smell the roses.

We already know that our government is throwing money and resources at opening up their data – why not open up the nation’s computing?

I know, it’s a long ask. And it’s for the long term.

But this is what we need to do – New Zealand is small, lean and mean. And if we can stay ahead of Brazil, India and China, we can make ourselves invaluable to their powering economies by being the testbed of web, cloud and network technologies.

So let’s do it for our economy – because you don’t sell cloud computing by talking with the technology guys and girls, you sell it by talking to the finance guys and girls.

An investment in a cloud that lets us design for the world and test and build locally, is an investment in our future greater than just laying fibre cable everywhere. Because the network has grown up – we’ve added the computing power into it and now it’s the cloud.