I was recently reading an analyst report on what’s left of the current Cloud Management Platform market, which caused me to smile (an infrequent occurrence)! It’s amazing how much water has gone under the bridge since the idea that one day all cloud computing capability would be equal and Enterprise workloads could be moved frequently to match where the lowest cost processing power was in the world that day.
Public Cloud for the Enterprise
At the beginning of this decade, Public Cloud technology was emerging for the Enterprise and the challenge of managing these growing number of “out-of-datacentre” capabilities started to raise its head. . This led to a technology known as the “Cloud Management Platform” – the idea that from a single software platform you could consume, manage and control your growing expanse of external Cloud capabilities. Enterprises could invest once in creating governance, policy and Cloud application templates, which could be portable between suppliers – protecting investment and avoiding lock in. On the face of it, all of this was a good idea and given the market expectation was for Clouds to commoditise over time, this would have been a valuable tool for the Enterprise IT admin or Cloud consumer.
Back then the Public Cloud market was different and the same as it is today different in that some of the names we know today were smaller players and some not even there at all. The same in that the services offered were all subtly different and difficult to compare like for like. Now if you think of insurance, as an example, there are many different types of insurers for your house or car but over time this market has commoditised to enable insurers to be compared on relatively few metrics. As a comparable example there is a huge market in insurance comparison, or brokerage, which is essentially the process of aligning your requirements to providers in the market and allowing you to compare based on the most important characteristic to you – normally price!
All of us early participants in the market had this understanding that the players in the Public Cloud market at that time would drop into four buckets in the next 2-3 years, this has proven true to some extent but perhaps not in the quantities we expected.
1. Verticalise – in order to avoid the race to the bottom in terms of pricing, many vendors were expected to lean towards a vertical and develop vertical specific propositions. This could be the likes of HIPPA accreditation in the Healthcare space or one of a growing number of certifications for the FS sector.
2. Specialise – an alternative to the vertical specific or pure commodity plays would have been to create unique offerings to offer alongside the base IaaS. This would be specific platform offerings – think Amazon Web Services capability which is only offered by them; e.g. Amazon Elastic Beanstalk.
3. Commoditise – this is where we expected the commoditisation to occur to enable like for like comparisons to take place. It was anticipated that IaaS would become a commodity market with standards for computing and storage with a standard unit of measurement for each to emerge. This would enable brokerage between vendors and price comparison trading to take place.
4. Die – if a vendor couldn’t achieve a degree of uniqueness through either verticalisation or specialisation and couldn’t compete with the global players in the commodity market they would quietly slip away into history.
What was predicted broadly speaking took place, vendors sought to create uniqueness amongst them in order to avoid being compared on pure price alone, but that created a problem.
Back in early 2012 I was invited to a launch event in the City of London based on the idea that commodity cloud capacity would in the future be traded like oil or coffee on a market exchange. In principle it was a good idea if the predicted commoditisation occurred and vendors would be willing to compete on price alone. The one key technical flaw was of course that it ignored the fact that computing capability is linked to data – and data is not, in the most part, free to move between cloud vendors at a moment’s notice to suit todays price point – at least not yet!
I guess the killer blow to this idea was the fact that the commodity collection of Public cloud vendors ready to fight it out on price never materialised. The true Public Cloud vendors such as Amazon and Microsoft sought to create uniqueness rather than create commonality – this put a stop to the constant race to the bottom in terms of IaaS unit pricing. This left the commodity bucket empty!
The Deutsche Börse Cloud Exchange launched in 2013 however quickly realised there was little to trade. Trading implies a degree a commonality in order to create a market in the first place.
The exchange closed in March this year.
Cloud Management Platform Market
The Deutsche Börse Cloud Exchange may have been a victim of the failure of the market to standardise IaaS services along commodity lines, however, the Cloud Management Platform market was also to fall victim to this failure.
The principle of an agnostic Cloud management Portal (CMP) was to broker workloads between cloud platforms, however, this would rely on the same assumption that cloud vendors would be offering the same services. Vendors quickly identified that any area of commonality would immediately be seized upon and draw them into a like for like price comparison. This arguably led to the creation of unique offerings as well as Public Cloud pricing strategies which were far from transparent or easy to understand (and compare) which I think has damaged the market as a whole – but that’s another point.
The resulting impact to the CMP market was that with little true commonality between providers any solution seeking to manage across all vendors would need to focus on only those common components– otherwise the principle of brokerage or managing multiple vendors will have disappeared entirely. This is what’s known as the lowest common denominator principle in that all of the unique values of each Public cloud provider would have to be ignored when consuming the service via a CMP in a bid to normalise capabilities. Naturally this wasn’t of much appeal to the Enterprise and the take up of CMP’s managing a multi cloud landscape has been minimal.
Where we are today
So back to the base of the story – the CMP market has now all but abandoned Public Cloud consumption and control and started to look inwards at the Private cloud market and has become largely swallowed up by proprietary vendors of Private Cloud software or hardware. Looking at the top six vendors all but one is now owned by a proprietary hardware or software vendor focussed on the non-Public Cloud market – (HP, Dell, IBM, VMware, Cisco). This will naturally lead to a product, which instead of focussing on commonality, will focus on the unique IP of their solution and form the control plane for these solutions.
It’s amusing that a technology that set out to remove vendor lock in the consumption of Cloud services will now become the principle cause of lock in to a proprietary vendor. It will be interesting to see how this technology emerges over the next 2-3 years.
I’m giving a talk on October 13th around challenges faced by the digital enterprise in a cloud world where these topics will doubtless come up – come along and join the debate!