he excellent article by BITKOM (Bundesverband Informationswirtschaft, Telekommunikation und neue Medien e.V.) is unfortunately in German (click here http://www.bitkom.org/de/themen/36129_61111.aspx), but summarizes the three current Cloud layers into Software as a Service (SaaS), Platform as a a Service (PaaS), and Infrastructure as a Service (IaaS). The article is one of the best down-to-earth summaries with actionable recommendation and check lists that I’ve read so far, but it does treat the Cloud as a commodity, as a plentiful resource. The danger, however, many articles from Mashable or ReadWriteWeb fall into, is to write about “unlimited computing power”. Throw in a bit of misunderstood Chris Anderson’s “Free” or “zero-cost of digital distribution” (while it is actually near-zero-cost plus a bunch of other smart things that people tend to quote out of context, or even worse, quote out of a review about a review of Anderson’s books), and you would have no market anymore:
Economics is the study of markets [...], the science of choice under scarcity.
Said Adam Smith, a pioneer of political economy, who lived from 1723-1790. So if we actually would have an abundance of computational power there would be no scarcity, and the economics of Clouds would end at this point.
ut let me take you back to early 2007, where I sketched the following timeline into my notebook:
From Thorsten's Notebook: The Cloud stack timeline, as seen in 2007
which is exactly the same picture I now – 2.5 years later – commonly use for presentations touching the Cloud space:
Cloud Stack by Thorsten Claus from 2007, now revisited 2009
2002 to 2005 – Web 2.0
Overview of Web2.0 service classes used in 2002/2003 to create an Internet TV station within three days
Mas 2002 I sat in a coffee shop with two friends during the three days following XMas and created an Internet TV station with 8 channels. At this time there was no talk about the “Cloud”. Everyting was Web 2.0, and Ismael Ghalimi [LinkedIn] started his Office 2.0 Database. We used about 20 or 25 Web2.0 applications, mashups, threw in some code, hijacked some services, and created the Station with about $300 CAPEX and $1,000 OPEX versus $30,000 revenue per month at its short-lived four month lifetime – before we took it down because of copyright concerns. We used some services that you would now call “Infrastructure-as-a-Service”, but at that time this was not called “Cloud Computing”, but rather “virtualized rack space”. But these things developed quickly, and Amazon started in 2002 famous “AWS – Amazon Web Services“… I’m not quite sure actually when they started to offer EC2, I think maybe mid 2006, but this is just a blog post, not a journalistic article, and I’m sure you can look this up on Wikipedia if you’re interested ;)
2007 – Cloud Platforms
n 2007 we the saw the first platforms emerging. People tried to get away from the red ocean of Web 2.0 services and from the discount commodities of computing Clouds. Instead, companies tried to offer more “intelligent” services and to become luxury item discounters. Cloud Platforms were born that were more than just yet-another-service, but rather platform enablers for better services. Here are some recent examples of these developments:
he list could go on, but I wanted to stress the different types of platforms that emerged between “simple” Cloud services and Cloud computing, creating a new complex ecosystem. And whenever a secondary market starts to emerge, every existing Cloud services and Cloud computing company has to ask itself whether its industry will become a commodity, much like gas or water or energy — and whether becoming a commodity is a good thing (because everyone will use your service) or whether you have to differentiate yourself with specialized offers to climb up higher in the value chain; basically serve both services and computing and platform elements disguised in any one of these layers. Examples for differentiation are:
2008 – Network Clouds and Relationship Clouds
y forecast might have been a year off, with major Twitter and Facebook gravity really taking off this year, 2009. A couple remarkable things happened. First, hardware vendors tried to jump on the Cloud band-wagon, beyond just providing computing hardware. They were joined by a couple of companies that were either not successful in creating gravity in their respective Web2.0 service, platform, or infrastructure Cloud layer, or who found these layers already being a red ocean. As a result, network Clouds emerged:
econd, a whole bunch of players tried to solve the identity puzzle. Most of them tackled the problem from the identity side – I provide you a service that you can use as identity. Of course there were (and still are) many of them. But even though Google also jumped in the game, the identity itself was nothing that was really valuable for users by itself – using one identity provider over another didn’t really make any difference. Early adopters started – oh paradox! – to use several different identity providers, some open (OpenID), some easy to use with pictures to click, etc.
It took a Facebook with its network of relationships tied to little social network centric applications to create the next Cloud layer of a “Relationship Cloud”. Relationship information and identity is extremely valuable in a land of two-sided business models and localization, and Facebook created an ecosystem of relationship-centric applications and services that makes it even more attractive for end users to use their Facebook ID to log into services than creating yet another username and password. I guess the reason why a “Twitter Login” isn’t taking off is the lack of applications around Twitter. Maybe not a bad thing, but let’s not get side tracked ;)
Other companies such as TuneWiki foster ad-hoc, informal relationships without explicit Identity to create value for end-users and their business relationships and advertisers, but they are still a relationship cloud player.
2010 and 2011 – Real-Time Information Logistics
he problem with all of these layers – from network Clouds to relationship Clouds – is that you have to make sense of all the information that you collect and process, especially if you have a two-sided business model. Moreover, the latest trends of viral videos will have an impact on your network and transcoding load, and when people hit your website, you would probably like to know what they were doing before that made them come to your site (more than just the HTML referrer header). It’s like Woopra’s “what’s happening now” dashboard, but think plus viral plus networks plus computing plus social. Plus “right now”. Now add M2M communication to the mix, and with it about 3.5 billion mobile info devices, 1,2 billion static info devices, 1,75 billion smart sensors, and 50 billion microprocessors and microcontrollers, and “right now” might be hard to achieve. Even with the “abundant computational power available through Cloud computing”, as many claim, it will be hard to make sense out of the information for various reasons:
As we can see, the information Cloud is by no means just a sub-problem of computing Cloud, because resources are not a problem, design is! All that comes back to the economics of the Cloud: We have a scarcity of well designed resources, and that will give us plenty of opportunities to create an economy around Clouds for a long time.
lot of companies are aware of this. Just search the VentureBeat deals for real-time information and you can see a whole flood of recent investments (Q2+Q3 2009). Factor in the expected average return on these investments, and you know that the created wealth in four years just out of the last quarter’s deals alone will create an industry by itself. Innovations reach from real-time information on mobile, to M2M, to website information, social networks, recommendation, location, local events, traffic, in-car Internet, …. If you address one single problem of real-time information logistics, you can probably get clients from 20+ industries, just as computing Clouds address many industries, or service Clouds address many industries. This is not a pure-IT-player game.
Even more, at the very “edge” of these innovations – quite literally – are large players like Cisco’s Medianet, Alcatel-Lucent’s Application-Assured VPNs, ArrisNet’s edge services, PeerTV, BigBand, NetLofig, Camiant, Procera, … the list goes on. P2P communication gets new attention because of the potential information and intelligence in the Cloud instead of expensive centralized data centers, where all information first has to be transported to as well, before even the first guesstimate can be taken. And in a few weeks you will read of a Tier 1 carrier, buying a peer-based distribution provider… Not because peer-based distribution is more effective than traditional Content Delivery Networks, but because this very Tier 1 carrier has understood the value of Relationship and Information Clouds that can be formed on top of its Network Cloud.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_c.png?x-id=75258564-5157-43bc-94e3-363dccd181cf)
