The rise, rise, rise of cloud computing

Cloud services sales continue to rise as user organisations find it useful to let someone else take care of the plumbing of enterprise computing. Suppliers are making huge upfront investments in – sometimes self-built – systems in large data centres and sweating their assets by signing up a rainbow of services types from software, platform and infrastructure. My Figure shows the growth of the three types of cloud services – Software, Infrastructure and Platform – ‘as a Service’. It demonstrates that the vision of suppliers such as HP was correct – we finally live in a period in which ‘Utility Computing’ exists, albeit as long as we’re prepared to let a modern, cut-down definition of the computing architectures we use are sufficient for the applications we want to build.

I don’t often write a lot about Software as a Service, because it’s a ‘no brainer’ for its suppliers. In 2018 why would any software supplier want to restrict its offerings only to programs which can be installed on ‘on premise’ computers? Nevertheless it is the largest of the three generic types of cloud services (see my Figure for the market shares of each in the last year). Since the outset Salesforce.com has been the leader; it has made good profits in the last five quarters, beginning to make up for the long periods of loss, which add to $805 million for the whole period from 2005 to Q2 2018. Microsoft, IBM and Oracle were the next biggest vendors in the year to the end of June. For users software has become more expensive as we’ve moved away from perpetual licenses to subscriptions and SaaS versions.

Infrastructure as a Service (cloud-delivered bare metal servers, storage and networking put into action by users deploying their own operating systems, virtualisation, management and applications) is the second largest of the three types. Its two largest vendors were Amazon (AWS) and IBM in the last year, with Google and Microsoft’s Azure services coming in in third and fourth.

Platform as a Service (cloud delivered resources including operating systems, virtualisation and management) was the smallest of the three types, but with similar leaders.

Regional market shares for cloud computing (see my Figure, where I’ve included just IaaS and PaaS services) differ – especially in Asia Pacific where Amazon has not yet been as successful as in the Americas and EMEA. In Japan suppliers such as Fujitsu and NTT Data became very successful following the earthquake of 2011 and have recently been joined by Chinese suppliers such as Alibaba and Baidu as leaders. IBM is the overall leader in Asia Pacific, with a 20.0% market share in the last year.

As a new comer to IT the rise of AWS has been amazing; in the last year it held a 30.7% share of these markets in the Americas (which are dominated of course by sales in the USA). In EMEA IBM was the leader, with a 17.7% share. It’s significant that indigenous vendors here are few and far between – France Telecom is the only one which appears in the top 8, although others, such as OVH, are making good progress.

Despite its continuous strong growth in adoption, cloud computing is far from being the most important element of enterprise IT (see my Figure for a comparison of its market growth with outsourcing/managed services, enterprise hardware (servers, storage and networking) and infrastructure software. It cannot be seen in isolation from these other areas – especially for large organisations in the modern world where regulations and security vulnerabilities are leading to  major investments in getting all of these elements to work together. We live in a multi-cloud world in which the use of AWS, IBM SoftLayer, Google, Baidu and other public clouds has to be integrated and managed securely into on-premise systems. Many users will develop new applications on cloud platforms, deploying those which require absolute security or give competitive advantages on-premise, leaving those which are not on an increasing array of cloud platforms.