Our CEO Nariman Teymourian likes to tease me about the public cloud. “You said everyone would go to public cloud back in 2010. This is 2016 and public cloud is still just 16% of enterprise IT.” HyperGrid is our third company together, and I like to think that we are getting better at them.
Our familiarity with one another breeds candor. “I did say that, and I was right — sort of.” Nariman smiles that smile. “I told you that it would be 15 years before the enterprise got to public cloud and it will be!” Nariman and I are talking about the future of the IT industry again. We are always talking about the future of IT.
Just a few years before Amazon AWS was created, Nick Carr famously declared that IT had become a commodity. Like the electricity power generation a century ago, Carr thought that IT would slowly consolidate into a few giant utility companies that provided IT as a service. Nick Carr was way more right than he was wrong. I hope that somewhere he has gotten a large tattoo that reads in bold letters “I told you so”. Carr wrote:
“What makes a resource truly strategic – what gives it the capacity to be the basis for a sustained competitive advantage – is not ubiquity but scarcity. You only gain an edge over rivals by having or doing something that they can’t have or do. By now, the core functions of IT – data storage, data processing, and data transport – have become available and affordable to all.” — Nicholas Carr, 2003, Harvard Business Review
Yet the transformation of IT toward a utility model has not proceeded as anticipated. Enterprises have not adopted public cloud as fast as predicted. Just a few weeks ago, JP Morgan released another report predicting that enterprise cloud adoption would grow by 20% CAGR for the next several years, with more than 40% of enterprise workloads running in a public cloud somewhere five years hence. What’s going on here?
Frankly, all of this talk about the growth of public cloud misses the point completely. The truth is neither here nor there, but somewhere in between. As one executive at Microsoft put it recently, cloud is a model not a place. The growth of public cloud is a symptom of the trend, not the trend itself.
And the trend, my friend, is software.
Software is eating the world. So it was written, and so it is true. But I think that few people understood the extent to which ordinary companies would become software companies themselves. According to Forrester, enterprise spending on custom software grew from just $50B five years ago, to more than $150B in 2016. Middleware spending shrank by $8B at the same time. Businesses are replacing off the shelf software with software they are writing themselves. They are doing this because it helps them to create competitive advantage. They can do this because there is an open source platform to stand upon. Building software has never been easier – or more complex – at the same time. Cognitive dissonance.
Let me give you a real-world example.
Last year I met with the CIO of a large healthcare system. He told me that Amazon’s GovCloud was highly attractive from a price and convenience standpoint, but the nature of his business made it impossible to even get started. His teams were building custom software, and they needed to do it within their own IT environment. I had to know why.
He explained that his software teams were writing code that did image processing for MRI. By doing so, their doctors are better able to detect cancer and other conditions than using the vendor supplied solution. From the moment the data is taken off the machine, the results have to be available to a physician down the hall in no more than 30 seconds. That was his SLA. Otherwise, you could delay a patient’s treatment, and that could have dire consequences. The speed of light is simply not fast enough to get the data there and back again. Not with current technology.
Here we can plainly see the reason enterprises are not adopting public cloud as fast as they should be — its impossible. Real infrastructure exists. Legacy systems still have to be maintained. Performance is paramount. Lives are on the line. Security is an absolute. Compliance is mandatory. Its a simple fact that complex systems like these are just not going to go anywhere fast. But they do need to change.
IT is complex. Over the years IT organizations became quite good at managing infrastructure complexity. Like a kind of factory, IT converted the raw materials of technology like servers, networking, and storage into integrated systems that delivered a business application. IT’s function in the enterprise was as a supporting cast member.
For 25 years IT was big infrastructure centric, delivering business critical services across a plurality of locations, users, and devices. IT controlled all of that end-to-end. For a little while, we could take it for granted. IT was everywhere all around us. In power grids and nuclear power plants. On oil derricks. Behind casino slot machines. These were happy times, but they would not last. The rise of the Internet and mobile computing has changed everything. The need to scale to support billions of Internet users is mandatory. The pressure to innovate to create competitive differentiation is immense. Today IT is not just a cost of doing business, IT is the business.
The fact that IT became good at industrializing technology is exactly what made it so resistant to the explosion of software. To make systems reliable, IT needed to control the combinatorics of software and hardware. Figuring out how to do that at scale is partly what makes infrastructure so time consuming. Changing any of that is what makes infrastructure so complex. Planning and acquiring all of that is what makes IT expensive.
What if there was a different way?
The early Internet companies struggled to scale their systems to support billions of users. Under relentless pressure to innovate, these companies needed to get changes to production as fast as possible. They could no longer afford to spend much time thinking about how to put the pieces together for every single change. Like Carr predicted, they began to think of their servers as commodities as opposed to custom built systems. We think that in the future all businesses will run their infrastructure this way.
If you cannot adopt the cloud utility, you must become the cloud utility yourself.
At HyperGrid, we believe that enterprise IT infrastructure is needlessly complex, and that developers are senselessly divorced from the IT operations process. Our goal is to make buying, building, and running infrastructure as simple and painless for IT as possible. We believe infrastructure should scale and integrate seamlessly, without months of planning, delays, upgrades, and downtime. We believe that deploying applications should be automatic, and that developers should help themselves. All without having to buy big infrastructure at huge costs — pay only for what you use. Its IT delivered to your data center as a service with all of the benefits of hyperconvergence. We call it IT at the Flip of a Switch™.
Welcome to HyperGrid. Plug us in, and turn IT on.™
James Thomason is CTO of HyperGrid.He writes frequently on cloud computing, IT infrastructure, DevOps, and software. He joins HyperGrid from Dell, where as CTO he led the vision, technology strategy, and product roadmap for Dell Cloud Marketplace, an online platform that makes it easy for businesses to compare, consume, and control cloud services. Mr. Thomason joined Dell through the acquisition of Gale Technologies in 2012, where as CTO he headed the product roadmap, architecture, and engineering of the company’s flagship cloud and converged infrastructure automation software, GaleForce. Prior to joining Gale Technologies, Mr. Thomason was the CTO and Founder of Virtiv, a San Francisco cloud automation startup focused on Linux virtualization, acquired by Gale Technologies in 2011. For over 17 years, as a specialist in distributed systems and large-scale infrastructures, Mr. Thomason has been an entrepreneur and innovator at a number of notable Silicon Valley start-ups, including Exodus Communications, Digital Island, Netli, NetVMG, Netscaler, 3Leaf Systems, and Ning.
- True Cloud for Carrier-Neutral Service Providers: Moving Beyond Direct Connect Models - March 7, 2017
- VMblog 2017 Virtualization and Cloud Prediction Series: HyperGrid’s Three #Predictions - January 3, 2017
- How the Role of CIO Will Change Forever as Custom Software Comes Back Home - November 18, 2016