The Coming Apocalypse of the Hyperscaler Advantage … and the Open Solutions Approach That Could Save the Digital Economy

At the ONUG’s spring conference in 2019, Dr. Douglas E Comer, a Distinguished Professor of Computer Science at Purdue University, spoke about a recurring phenomenon in technology deployment: the lack of system interoperability that limits economic growth for businesses. 

As a member of the original IAB, he participated in the early development of the Internet and is internationally recognized as an authority on TCP/IP protocols, his expertise and experience provide a backdrop for insights into the future of cloud computing. Here is a synopsis of his 2019 talk.

Let’s Go Back in Time 

An excellent example of what I mean by system interoperability and its impact on business growth is the rise of TCP/IP standards. When businesses first started using TCP/IP, each vendor had a different solution with no motivation to work towards an open standard. How did the situation change? Customers. Businesses came together and refused to buy computers or network boxes that didn’t support an open internet protocol. This same pattern is occurring in the world of cloud computing. No interoperability exists among cloud providers; yet their revenue continues to grow. 

The Pattern

Perpetuating this pattern has a lot to do with sales tactics. I found that cloud providers sell CEOs on the concept of cloud computing by telling them they are paying too much keeping it in house. They explain the high cost of equipment, maintenance, expansion, and personnel. Then, they suggest outsourcing. They promise cost savings and a painless transition. They explain how they can do it better and less expensively because they can achieve an economy of scale. 

If you look at the hyperscalers, you’ll notice that cloud computing wasn’t even on their radar as a revenue stream five years ago. Today, it is 29% of Microsoft’s revenue and 11% of Amazon’s. According to Microsoft, they spent $20 billion setting up Azure. $20 billion. What did that money buy?

They needed hardware, but $20 billion buys a lot of switches, routers, and computers.  That money went to create new services. Services that they are now selling to the same CEOs who migrated to the cloud. These hyperscalers are selling companies “tools they could never have built themselves.” They are continuing to lock organizations into their cloud solutions.

The Strategy

These sales tactics are part of an overall business strategy. The hyperscalers want to lock as many businesses as possible into using the cloud. They want these same businesses to purchase their services until it becomes more cost-effective for organizations to pay a monthly subscription fee. That fee may include products and services that companies may or may not need.

This business model has a downside for the hyperscaler as well. When they lock-in customers, they have a set of customers that can’t easily change. If you ask mobile phone companies, they will tell you there are really very few new customers. Mobile providers are simply stealing from each other, buying out contracts, and hoping that the number of stolen customers will balance out the contract buyouts. 

Hybrid solutions, when you think about them, are no better. Companies can choose from two options, but they each have their limitations.

  • Companies purchase public cloud solutions, make their private cloud, and manage it. The problem is a company is tied to that provider’s solutions.
  • If companies decide to purchase from multiple providers and place the cloud on their premises, they experience a lack of interoperability first hand. Sometimes one solution can’t cross the cloud infrastructure. Finding a solution may further limit a company’s flexibility.

In these scenarios, companies wait for the next disruptor who will break the cycle and give them options. Eventually, the disruptor becomes part of the pattern that repeats itself.

The Disruptors

Where can companies look for possible solutions to the interoperability issues? Where are the disruptors? As outlined previously, it doesn’t make business sense for companies to try to replicate a cloud solution. It’s too hard to compete economically.   But what about the research or open-source communities? Can’t they help find a more open cloud?

Research Organizations

I visited a few research universities throughout the country, talking to them about the current cloud computing landscape. I asked them to take up the challenge and build an alternative. Sadly, the academic community was not interested. 

Open Source Groups

I moved to the open-source groups. I presented the landscape again, but no one was interested. The groups were only interested in building an SDN controller or a platform.  They were not interested in data analytics or data science. Even though groups such as OpenStack say, “Do not use a proprietary cloud solution,” don’t expect a solution from the open-source community any time soon.


Thinking back to TCP/IP.  What would happen if companies went together and demanded cloud interoperability? As customers, it’s not your responsibility to tell providers how to do something. It is your responsibility to tell them it needs to be done.  It is also your responsibility to use your purchasing power to ensure that changes happen. Without that coordinated effort, companies will continue to be restricted by proprietary systems that turn into legacy systems that limit technical innovation.


What I tell my students as they graduate is: You have learned how things are done, how things are built, and how things work. Now, you need to imagine better. Stop thinking within the box and think about making the world better. For businesses, I suggest the same thing. Think outside the box. Be innovative in the solutions you need. Then help organizations share your vision so they can make it happen.

Join us (for free) on May 6-7, 2020 for the ONUG Digital Live  to learn how different industry sectors are using technology to digitally transform their organizations into innovative cultures with visions of the future that are outside the box. We hope you join us!

Author's Bio

Dr. Douglas E. Comer

Distinguished Professor of Computer Science, Purdue University