PaaS moves on with Cloud Foundry

Yesterday, VMware announced the launch of Cloud Foundry. This is a big step in the evolution of Platform as a Service.

What is profound, is that this allows you to build applications in a private (your data centre), public (Amazon’s or Rackspace’s cloud) or hybrid environment (a combination of both). It scales from your laptop, to the hundreds of servers in the data centre.

Finally Cloud foundry is Open source. This is important to you guys out there because this should in theory allow you federate (or connect) clouds together. The theory being that as you have a known code base and API set, you can readily do this task.

In the cloud space, this is big news. PaaS is going to be the most critical enabler of Cloud applications going forward, but todate the execution on this vision by potential leaders like Microsoft (Azure) and ( has not quite been there…

For some more coverage of this move check out

Krishnan at

The Wikileaks Amazon saga spotlights clouds Achilles heal, data sovereignty

Disclaimer : This is not a post on the right or wrong of Wikileaks, this is about the implications on public cloud computing.

Last weeks events around Wikileaks and Amazon are a startling reminder of just how different the public cloud computing model is when it comes to data law and responsibilities.

As i’ve highlighted in the title, to me this clearly shows one of the Achilles heals of cloud computing. Some don’t agree, but those some are mainly based in the USA. See some quotes below

“It’s USA-centric because it’s analyzing a USA-centric issue: U.S. govt cables, a US data host, and US law jurisdiction over that co”

My perspective is international.  Here’s how I see it (and this is high level).

  • An international organisation chose a US ’tied’ cloud provider to be their IaaS service provider
  • Some folks in the government took a disliking to their activities (in this case the content) and reportedly leant on the cloud provider
  • The cloud provider stopped providing service.

Now I acknowledge that this is an extreme case, but at its most basic level the same could be true of any international organisation. For example

  • I create a company that has a product that attacks a core US industry that has millions or workers and lots of revenue and tax income (Finance, automotive, pharmaceutical, IT, whatever), I choose a US based IaaS provider (because its large and costs the least)
  • The industry feels the pinch and lobbies the hell out of the senate (or the house) to do something about it because lots of jobs are on the line etc etc
  • The government intervenes, protecting its national interest (and that happens all the time) and has my company terminated from the cloud provider
  • It is also possible that they might go after the data (which incidentally violates my national data protection rules ). What then happens to my IP?

I’m not trying to scare monger here. What I am trying to do is point out a few salient facts about the international dimension of cloud computing. Specifically:

1)      Cloud computing is an international phenomenon, so many users will take services from a cloud provider who is governed by different laws, politics and sentiment.

2)      These users of cloud services operate under different data rules, and what may be legal or commercial or private in their nation could collide with the cloud providers  national laws

3)      The political machinations of the hosts government whims can have major impacts on the user, industry and trade

4)      The concerns voiced over data sovereignty over the last few years are actually valid

To me the Wikileaks case is a clear pointer to greater concerns about data sovereignty within the cloud community and a continued drive towards regionalisation of cloud.

Ok, So what do I think the future of cloud computing looks like?


My last post on what I thought the future of cloud computing looked like didn’t evoke the commentary or debate I had hoped… one of the problems with blogging I guess.


In that post I asserted that the similar industries had already given us clues as to what the future of cloud computing looked like.


“.. [look for] historical likeness, where have you seen this pattern before?  Rail, roading, electricity, shipping, telecommunications, the internet, grocery stores, banking (particular the US)…”


So ignoring the twists and turns of industries that represent them making their way, what did I mean when I think they show us the future?


They all show similarities (at the highest levels).


Examining the history of rail, banking and telecommunications , they all adhere to similar phases.


  • The fragmented start up – where early investors see the opportunity and build out an offering usually in one geography (ie county or state).  The most important thing to note, is that this pattern re-occurred in every nation. So not only multiple national hubs occurred, but every nation had hubs


  • Arrival of standards – users demand saw players interconnect because there was more value in this than just one provider – rail it was gauge, telecommunications IP, banking networks…


  • Emergence of dominant players – these systems all benefit from scale, so natural competitive forces lead to large players winning and small players loosing


  • Monopolistic threat or behaviours naturally occur after or during the dominance


  • Government intervention and regulation – to appease the consumer but also to control a strategic asset. These systems became so important that control became inevitable…


So what does this mean for cloud? I think we will see a similar pattern.


  • The Fragmented start up – CHECK

  • Arrival of standards – well people are calling for it and there are some

  • Emergence of dominant players – CHECK, I predict a lot more M&A activity in the near future . This will mean that all the mainstream cloud (is there such a thing?) being done by a small group of core providers with all the cool, innovative stuff begind done by small outfits and happening at the periphery,  
  • …who do I think will be the core providers?  SFDC, EMC/ VMWare, Cisco, Amazon, IBM…I’m struggling to see MS making it “you cannot disrupt yourself” according to the innovators solution


  • Governmental intervention – well that’s either inevitable or being asked for already


The bit that isn’t being seen yet but I believe will come is regionalisation. But when governments realise that cloud ‘could’ mean that a lot of IT jobs are threatened, eventually they will get to the conclusion that local clouds will give them most of the outcomes and retain employment…. and that will drive regionalisation.  I will probably by couched under some moniker like data sovereignty, but that will be the most important reason….


So there you go, that’s what I think the future of cloud computing looks like…

The internet, the great business facilitator.

I’ve just read a fascinating article by Tom Foremski  called “ The internet devalues everything it touches”. It raises some great thought provoking points, it’s also slightly off base.

Let me explain. The internet is what it is, a massive and cheap to use platform that connects together many millions or billions of users.  The value destruction component isn’t a function of the internet, it’s a function of how people are using it to disrupt entrenched business models.

If you look at the examples mentioned by Tom, they all relate to some form of distribution efficiency. The simple fact of the matter is that for digital assets, the internet is more efficient at moving stuff around than roads, rail, shipping or retail stores.  The take away is, that if you are involved in traditional distribution and the asset you move can be digitised, you are in trouble. You will be disrupted.

I think it is important to note that it doesn’t have to be value destroying, check out what Amazon has done to the book value chain. Sure they’ve brought down the price, but more importantly they’ve removed stages in the value chain and absorbed their contribution while eliminating their cost.  In effect they’ve made book retail more profitable, and because they did it, they took a larger piece of that profit…nice work if you can pull it off.

Amazon knows what its business is about

“The Business of Delivering Stuff

Amazon is a master of the supply chain”

That’s why they got into Kindle. The digitisation of print, combined with the (nearly) free distribution of the internet is in fact the perfect business model. It so grossly simplifies the distribution of the asset, that the profit pool expands even though the revenue might contract.

There are other examples of this in action. Many look like companies buying up companies to create vertically integrated solutions, but I think it is more about optimising distribution COSTS.

Level 3 bought the Content Distribution business of Savvis, which on the face of it looks like a vertical play. But when you look at who Level 3 already service (Youtube in particular), this looks less and less like a growth play and more like a play to optimise (content) distribution and its associated cost. I suspect that Level 3’s business case would have stacked up on the cost out alone and discounted any upsell!

Google today announced its intent to buy On2. Why? Well in the Youtube value chain Adobe flash became a cost liability (both in terms of licensing and processor). Google already has the massive distribution of the internet going on, so what do they do? They vertically integrate by acquiring an asset to streamline out some cost. The internet didn’t destroy that value, all it did was facilitate distribution.  Google can choose to absorb the cost as profit or pass it through.

The media companies will probably disagree with the above example, but then their historical value was in distribution wasn’t it?

To me this is just disruption in action, in this case the internet is the facilitator of the low cost business model.

How to make money from SaaS #2

This is the second part in a series on how to make money from SaaS. If you missed it, here is Part 1 (It also appeared on Cloudave here)

Step 4 – Create a plan

This is the critical step in the metamorphosis from idea to actual start up. Why? Because it's the bit that describes the "how", and that information is really really important to investors.

No matter what your product does, you are going to need money to make it real. If you can articulate the process by which you are going to build and operate the product. How you are going to sell and market the product, how the business is going to run, and hence how you are going to turn a profit, you are in a good position to get funding. IPO, private Equity or Angel…. understand which is best for you and go with it.

The plan will also help you focus your staff, there should be one clear message that you can tease out of the planning process. 'Get Big Fast' (Amazon) or 'A PC in every home' (Microsoft) are good examples.

Step 5 – SaaS companies should buy SaaS

For goodness sake use your own software, and take SaaS every other way you can, bar one application (and I'll get to that). You have to live and breathe this, experience the highs and lows from a customer point of view. That way you keep your upfront costs down, know first hand what a pain in the butt an outage is, and understand interoperation and single sign on from a customer point of view. I repeat Live and breathe it.

But, keep one on premises application. For a similar reason, so that you'll understand the highs and the lows and have a balanced view. Your sales people can use this… "we deliberately kept our XYZ application on premise so we'd remember…"

Step 6 – Think long and hard about how you will build your service

Sinclair Schuller of SaaS Blogs (and CEO of PaaS provider Apprenda) wrote an excellent dissertation on the many acronyms of cloud computing platforms. The key here is speed, economics and reliability.

I refer you to point 5 above. Be a SaaS company. In my opinion, there is no way you should build every element of your service, that would be fiscally irresponsible given the plethora of platform offerings available today.

This opinion is especially true of the underlying infrastructure that delivers your service. Think about what goes into building a geographically diverse, robust operating environment. You also need to think about how and where you are connected to the internet because it can significantly affect the end user experience. All of this costs and absorbs time. Why even do it? A good PaaS provider has also thought through such things as billing, API interoperation ( and Amazon) etc. All very important, and sometimes easily forgotten.

If you are going to build stuff you have a lot of different choices depending on what you have as a service

  • Build it all in house – do this in a proprietary system or opensource
  • Go PaaS and only build the 'special sauce' part.
  • Go IaaS, build the OSS/BSS layers plus the special sauce

Any approach you take is going to have its pros and cons. To me the most critical part is to understand your costs and cash burn rate. Then look really hard at the numbers and seriously, in your heart of hearts, ask…. can we make money doing it this way. If the answer comes back positive go for it, if the answer is negative or marginal, consider the alternatives above. proves that actions are better than words

I’ve just had the opportunity to catch up on the events of Dreamforce, I must say i am surprised about the lack of attention that their announcements are getting in the blogging community. I dent get it because all of their initiatives are at the forefront of cloud computing. They also have significant ramifications for their competitors. (incidentally who ARE their competitors is a interesting questions.

Lets quickly recap what was announced. Firstly a mashup with Facebook. Secondly integration with Amazon so that SF.DC becomes the orchestration layer to E2C. And the Google Apps integration.

What’s really important to understand her is the evolution of SFDC business model. The facebook ( & LinkedIN) tie up allows them to move from a B2B business to a B2C. They are effectively providing their infrastructure policy & orchestration layers to their customers. This does several things. It makes them very sticky. It drives enormous scale. It gives them access to the mother of all Identity stores. Most compelling is that the viral nature of the social web enables them to make you, the FB user their headhunter, lead finder or product evangelist. It makes them a huge transaction hub , and once you’re there you effectively cant be displaced. The reason for that is you become the centre of gravity (or you get to dictate terms!) check this out for much more detail.

The second powerful theme is the evolution of PaaS. This is a much more mature play in that it effectively leverages the AAA, policy , rules and control layers that SFDC invested in for their own platform, to provision services in other cloud platforms. This is clever for a number of reasons. Firstly the customer is owned by SFDC in so far as they are the primary provider. Secondly it effectively relegates and controls where the other cloud play. Thirdly they *could* become the default start position for cloud services as ubiquitous as ‘Google’ is when you think of search.

This is a massively powerful position that SFDC are aiming for. If they are fast (and this is a race strategy) and don’t scare partners too much with the degree of control then this is a ambitious and winning play unfolding…