The Gens Law, or why public cloud will be bigger than private cloud
Follow the apps ...
Published 10:27, 09 July 11
At last week's 2011 Cloud Leadership Forum in Santa Clara, one of the highlights was the opening session on 12 key things for CIOs to understand, on their road to the cloud.
A hot topic inside and outside IDC has been the long-term prospects for customer premise-base "enterprise private cloud" implementations, versus public cloud services - how customers will choose to spend their IT resources, whether private cloud is more than an extension of today's traditional customer-run computing environments or not, whether public cloud is really transformation, and will be widely embraced for an increasing set of workloads and services, and whether the "wall" between the two will become far more porous, as solutions have built-in hybrid characteristics.
One contentious prediction was made by Frank Gens, who stated that "public cloud will be more important than private" to customers, to ICT providers, and to solutions developers.
The notion is that we're just starting to see the outlines of a Darwinian explosion of deployment options and topologies with cloud, but ultimately, public cloud services will be more pervasive than private clouds because most new solutions will be developed for public cloud platforms as public clouds offer the most favourable distribution channel for ISVs — and users will follow the solutions.
Public cloud SPs will have significantly lower unit costs than many private data centres, making it tougher for private data centres to justify retaining systems in-house where there is no compelling argument otherwise (e.g., security/privacy and custom/differentiated solution).
Indeed, the largest public cloud SPs will provide a higher level of security and availability than even the most sophisticated private data centres because they have greater resources (as shared services providers) than most customers. And the functional and trust gaps between private data centres and public cloud services will narrow greatly — and in many cases, reverse — as most of the industry's entrepreneurial energies (and investments) are aimed at mitigating public cloud adoption inhibitors.
So with more than 30 years of working in the industry under his belt, Gens' simple but powerful law is this: “Markets always follow solutions."
This pattern has been repeated before in the IT (and other) industries.
It is part of the "disruptive innovation" phenomenon identified by Clayton Christensen, and is a foundation for the disruptive economics of the cloud model, which is fuelling the aggregation of the industry's economic and innovative resources around public cloud platforms.
A look in the rear-view mirror gives us a frame for how to think of the next ten years. The employment, workload types, and configuration, of servers has been variable since the development of server architecture in the 1980s.
Under a relatively stable revenue topline, there was ongoing change happening in the reflexive relationship between servers, PCs and packaged software, the one driving and changing the expansion of the others.
In particular, the change in supporting architecture (especially the big surge in x86) entered a "virtuous circle" with a big change in software: independent solutions vendors could see the trendlines emerging and started to skew development efforts toward x86; as a result, users saw more/better software on x86, which drove more demand for those servers, and so on.
IDC sees the same thing happening now with the "3rd Platform" of public cloud, and the development activity surrounding solutions development.
IDC has built into its assumptions for public cloud a substantial "apps effect" boost, in that the acceleration in public cloud adoption, driven by the growth in applications developed for the public cloud, should kick in by 2014 - as the major cloud app platforms/ecosystems (Azure, App Engine/Marketplace, Force.com, Cloud Foundry, etc) reach critical mass with developers and solutions.
This assumption is based on strong historical precedent - over the last 20 years, server and software investment skewed strongly toward the Wintel world because of:
- Better economics
- Key vendors adding more and more enterprise capabilities to the new environment, and
- Developer (and new app) migration
The importance of software developers, and where they choose to focus their energy and target new apps ecosystems as ripe for innovation and profit, is huge: the PC was introduced in 1981, and from 1980-1986, the packaged software industry grew 10X, from $2B to nearly $20B, with a lot of that growth around the PC.
Beginning in that period and continuing to today, we've seen a refresh of existing infrastructure that drove a stable baseline of server demand, but a commensurate, and even larger, expansion of apps and users.
Thus thinking about the history of the last, second platform (client-server and PC), gives us the template for thinking a little more specifically about where growth will come from in 2011-2015.
Private cloud systems will undoubtedly take a significant portion of the refresh/replacement of existing IT systems - as an evolutionary path of existing on-premise IT environments; so we see its growth tracking the system replacement rate, plus a minority portion of new apps.
Public cloud will also be an option for some replacements (already seeing it with email workloads, e.g.), but we argue that it will be a less comfortable option than private within the 2011-2015 period, and will grab a minority of replacements.
But when it comes to where new apps (and therefore most of the industry's growth) will go, you have to look at where developers are likely to create and deploy those apps - and the value proposition for developers of public cloud development and deployment is extremely strong.
Back to Gens' law - "markets follow solutions"
Last year (and this year) IDC called 2014 an inflection point in public cloud growth - a "knee in the curve" - where public cloud development platforms (both "PaaS"-like test and development IDE platforms, and marketplaces built around these platforms) mature, and developers skate to that new cloud "puck" in search of low-cost developer tools, and access to customers.
The result of this growth accelerator will be the build up of public cloud apps, the development of cloud-savvy channel partners, and the improved cloud-readiness of emerging markets.
The 2010-2015 growth rate of public cloud (see IDC #228485, 2011 Public Cloud Forecast) is growing at 26%, where public cloud spending as a function of all IT spending in IDC's Black Book is still less than 20%, but cloud is a driver for more than 30% of incremental growth, just from 2014-2015.
IDC predicts that more than 80% of net new companies coming to life in 2011 will be operationalised around creating, testing, and provisioning a cloud service, rather than a packaged solution.
Clearly, cloud is a huge net driver of IT market growth, and developers are creating solutions which map to that future.
Posted by Robert Mahowald
cloud, platform, PaaS, x86, server, growth, law, Gens,