Cloud 2.0 Tornado: Three Predictions for 2017 and beyond

Cloud 2.0 is here. The public cloud market has crossed the chasm. Whether they realize it or not, all the tech players are entering a tornado.

Cloud 2.0 will be the proving ground for the next tech giants of our industry. Which companies will win the rich rewards of the massive mainstream market? Which ones will join ranks of the dead or also-ran companies?

As we welcome 2017, ’Tis the time of the year to make predictions, so here is my perspective how this will play out in the next 3–5 years.

1. Long live Public Cloud

Yup. We are entering a future dominated by public clouds. By various estimates, 50% of all customer workloads will move to public cloud in the next 5 years.

Will Amazon Web Service (AWS) continue to be the undisputed leader of public cloud?

It seems that way, right? Especially when we observe how AWS left a blood trail of public cloud victims in its wake: HPE Helion, Cisco Intercloud, Rackspace Verizon/Terremark, and CenturyLink/Savvis. Significant high-tech fortunes were lost as these companies tried to cross the chasm but found themselves in the ditch.

But the rules are different in the Cloud 2.0 tornado.

This is the game of sustaining engineering, incremental innovation, and catering to enterprise customers who care more about predictable business outcomes than about playing with the latest tech toys.

This is a game that existing market leaders are extremely good at.

Large software and services players, like IBM, Microsoft, and Oracle, however, need to fight hard to win this market. They should prevent existing customers from fleeing while winning net-new customers. But in order to stay in the game, the incumbents should rapidly accelerate their transformation from the old model to the new model: this requires a bold, top-down executive and board-level commitment, reallocation of resources, and refreshed sales and business processes.

If they get their strategies right and execute well, IBM, Oracle, and Microsoft can mount a significant challenge to AWS dominance in the next decade. Microsoft is already well on its way “cloudifying” its existing enterprise customer base. IBM and Oracle in particular have full stack (IaaS, PaaS, and SaaS) offerings but they have a long way to go to catch-up with AWS. If they integrate their full stack solutions well they can offer a seamless one stop-shop customer experience. Additionally, because of their hardware strengths (POWER and SPARC respectively), IBM and Oracle also have the potential of innovating at the silicon layer to offer differentiated bare-metal, IoT, Machine learning, Big Data, and container solutions with better performance and security at lower costs.

Google is another major contestant in this battle. But unlike Microsoft, Oracle, or IBM, I think their biggest challenge is their relative lack of history or track record with Fortune 500 enterprise business, a market which is markedly different than the mass consumer market that underlies Google’s core DNA. It remains to be seen how Kubernetes and other moves such as integrating Cloud Foundry into Google cloud is going to play out. It’s too early to tell.

The battle for Cloud 2.0 dominance is going to be fierce as market enters the highly rewarding tornado phase! While it’s conceivable that AWS will maintain its lead for some time to come, it’s hard to predict other winners. It’s all up for grabs. The incumbents who are able to boldly cannibalize their existing cash cows and invest heavily in the new model stand the most chance of making it through the tornado.

2. Private cloud is dead, long live private cloud- “as a service”

Most enterprises found out the hard way that building a private cloud involves complex integration of dozens of software and hardware components. They realized that keeping all moving parts in sync and operating smoothly and reliably at scale is a hard problem. They simply didn’t have the skills or knowledge to cost-effectively build, integrate, and operate complex private cloud services. Nevertheless, the vendor community continued to sell hardware, software, and integration services forgetting that cloud is all about ongoing 24x7x365 operational services that needs to be backed by business SLA’s.

This is one of the main reasons private cloud market never really took off. I offered a few other explanations in my previous blog post: Private Cloud is dead.

So then, is there another private cloud model that can work?

Sure there is.

Let’s call it Private-Cloud-as-a-service for lack of a better term. This is how I would define it:

A fully vendor-managed (hosted or on-premises) cloud service that offers a public cloud-like experience in a dedicated and secure data center environment so customers don’t have to deal with the complexity of building and operating their own private cloud.

There is also a compelling cost advantage — sometimes up to 50% cheaper — compared to a public cloud for most mid-sized deployments (<2000 servers). These solutions should appeal to cost conscious customers with security, privacy, & compliance concerns and those who simply don’t have access to public cloud services which is particularly true in Europe and Asia.

Rackspace was amongst the first to recognize this problem. They took their company private in order to accelerate the pivot to managed services for not only hosted and on-premises clouds but also public clouds like AWS and Azure.

In addition to offerings like Cisco Metacloud and IBM Blueblox, a number of mostly Silicon-valley startups are also quickly jumping into this game of providing Private clouds “as a service”. Here are a few examples:

Mirantis: Build and Operate an on-premises OpenStack cloud
Platform9: Run OpenStack or Kubernetes as SaaS (hosted control plane)
ZeroStack: On-premise cloud as a service fully managed using cloud-based monitoring & operations
CloudSimple: Cloud infrastructure service (hosted) for enterprise applications
StratoScale: Delivering on-premises AWS region and Kubernetes-as-a-service

This market is relatively early (compared to public cloud) and is yet to cross the chasm. But its also a massive market — think of it as the rest of the global market that public cloud will not address in the next decade.
Every player in this market has a shot at winning by addressing different use cases and workloads (e.g. DevOps, Big Data, and IoT). More than likely, companies will use open source technologies (Linux, OpenStack, Kubernetes, Docker, modern DevOps tooling etc) to build out their platforms.

But customer could care less what technology (open source or otherwise) is being used under the covers. All they are looking for a reliable service with business SLA’s.

So, this type of business takes a different kind of staff and operations to pull this off. The model is very different from building software or hardware and shipping it. This is about site reliability, monitoring, life cycle management, and running an operations center to proactively maintain customer SLA’s. And there’s always the looming threat that public cloud will eat their lunch!
I think Rackspace has a good chance of leading this market since hosting & operations support is in their DNA and that is what it takes to run a private-cloud-as-a-service.

3. Hybrid cloud is a myth. The real Hybrid Cloud is here (finally)

If the world is going to be dominated by public and private-clouds-as-a-service, it’s only natural to think that customers would want to use these together. The so-called hybrid cloud that people talked about for the past several years is a kind of hybrid cloud where any private cloud will work seamlessly with any public cloud managed seamlessly through single management/control plane.

Well that never really materialized despite many CMP and hybrid automation tools.

Why? My previous blog post explained the reasons in detail: Hybrid Cloud is a Myth.

I will repeat here what I think a real hybrid cloud is:

It is one where the on-premises and public environments have been pre-integrated/pre-packaged or alternatively architecturally compatible and managed by a single provider who can provide industry SLAs.

This market will be dominated by vendors who will bring architecturally compatible hybrid offerings such as Microsoft Azure Stack (works seamlessly with Azure public cloud) and Oracle Cloud Machines (exact replica of Oracle public cloud) or Pre-integrated and pre-packaged hybrid solutionsofferings such as VMWare on AWS and Salesforce on AWS. Startups mentioned above ZeroStack, Platform9, and StratoScale all have hybrid offering with AWS.

A few parting thoughts

I can’t help think about how broken and dysfunctional the enterprise software model is today and how mindlessly complex and utterly inefficient a modern day enterprise data center has become. It’s just a large amount of effort and a waste of time deciding what technology to use, what software to buy and so on. Isn’t our time better spent in helping the needs of the business and let the experts provide the services we need?

I think enterprise data centers will eventually disappear (it may take a long time) and everybody will consume digital “as-a-service”. If nothing else, our millennials who are already “spoiled” on a diet of real-time cloud services on their devices today, will the agents of massive change in the enterprise over the next decade.

It’s all good though. The most important thing is to never to stop innovating. This is how we can make the world a better place to live for generations to come.

I wish everyone a wonderful and fulfilling new year ahead.