Cloud Computing Predictions Are Over-Hyped
Ten years ago we saw the rapid emergence of the content delivery network (CDN), and as is the case with many "new darlings," we saw the valuations overinflated and based on no real understanding of the market size and how quickly it would emerge.
Now, a decade later, CDN has settled down to being a commodity product. Prices are skinnied down to minimum margins. Differentiation is by logo, value added service, and relationship, and not by fundamental feature set. Yes, some get to market with new formats and great new features first, but only stay at the front of the pack for a few weeks before they are joined by their peers.
Over the past few years we have all seen virtualisation take off, and volatile computing, virtual hosting and various manifestations of software-as-a-service, platform-as-a-service (PaaS), and more recently infrastructure-as-a-service (IaaS) computing have started to emerge. There is no doubt that the nature of infrastructure and shared resources has changed utterly.
For those of us involved in workflow and distribution, this is what we think of when we think of "the cloud."
Having been closely involved in the marriage of cloud and CDN over the past three years, I have a clear picture of what cloud means to me. But now I see everyone—and their grandmother, their son, their three cousins, and a few mates from the pub—talking about cloud-this and cloud-that. So much so that I am beginning to realise that in the post-iPhone generation, ANYTHING that happens remotely is being termed ‘cloud’.
And so we see Cloud services getting hyped. Not just hyped, but so over heated that I think they face problems.
Let me throw a few thoughts at the wall here:
- The main driver of broadband is video.
- Video makes up more and more (certainly eventually the vast majority) of the traffic on the internet.
- The main delivery mechanism for this traffic is the CDN.
- The CDN market is worth somewhere in the region of $1 billion.
- So the vast majority of the world's internet traffic is delivered by a group of mature companies who are worth in the region of $1 billion.
- Amazon, which owns the only true global IaaS infrastructure that is widely available as a true coud have a total market cap of $100 billion. This includes all their services, not just their cloud offering.
- Yes, Rackspace and GoGrid are global IaaS cloud providers, but there are no other really scaled out global IaaS cloud providers, and if you really try volatile compute with Rackspace and GoGrid you will find they are really just virtual hosting providers—they have some way to go before they can give ad-hoc infrastructure on a short notice high availability to the same degree that Amazon does.
- In fact, Amazon is estimated to be generating less than $1 billion annually from its cloud computing services. This is growing rapidly, but this figure is probably about right.
So when I saw a press release from Martin Hingley's research and consulting firm ITCandor that stated “the UK cloud computing market will be worth £119 billion by 2013," it caught my eye. The first thing I did was checked in on the UK’s current gross domestic product, which is £2.175 trillion. So this chap Hingley reckons that in two years' time the UK cloud market will be worth 5% of the UK’s GDP.
Please give me some of his Kool-Aid.
I have read his slides and he does throw in absolutely everything that could be related to clouds – from telecoms to ISPs to servers, and on and on.
And this is the problem.
There is now an audience out there who seriously think that the UK cloud industry is going to be worth a hundred billion. It's this type of absolutely ridiculous overheating of a sector which will cause over-investment, a reputation of failure, significant market corrections, and huge setbacks in the mid-term.
The fact is that, for a large client, I searched high and low for a backup to Amazon EC2 that had strong Europe presence. This was for a highly demanding computing environment where we may requisition 300 to 600 servers for use for an hour with less than 30 minutes notice, and we may do this thousands of times a year. Amazon EC2 handles it fine. There is no other provider in the UK that would even pass our initial tests. None.
My client’s spend on this works out in the single figure millions per year. It is as big a client as these clouds have seen.
So how this then scales the market out to £119 billion, I really really really don’t know.
And the problem is that if cloud gets so hyped that we all think its more than it is, we are in serious danger of disappointing many people, on many levels.
In the Streaming Media world we know that the hype results in attrition, and many CDNs have come and gone. Many great business models were lost because the opportunity was overinflated and investors' and entrepreneurs' expectations were simply not realistic at the outset.
The cloud community, which is not a million miles from the CDN community (in fact, several conferences have a CDN track and a cloud track co-located at the moment) would do well do stop and reflect a little before it seriously goes off the deep end.
More patience, more reality, less outrageous claims, less bull**** statistics, and more stability, please