Multi-Cloud Is a Losing Strategy: Here's Why

The world has changed a lot in the last ten years. Today, every company has a “Cloud Strategy.”As Cloud Experts, we no longer have to put expansive efforts into evangelizing Cloud. One thing that has remained a constant since 2009 is all of the hoopla around Hybrid Cloud. And I’m here to tell you: Hybrid is a Myth. At best, Hybrid is a stepping stone on the way to an Enterprise being completely Cloud native. At worst, the Fear, Uncertainty, and Doubt (FUD) campaign launched by legacy hardware manufacturers amounts to a last-ditch effort to save their companies by flat-out lying to their customers. Either way, it’s time to put the tale of Hybrid to rest once and for all.

The Beginnings

When a disruptive technology like Cloud Computing emerges, it forces affected industries to execute on an adapt or die strategy. In the last ten years, we have seen significant pivots by traditional hardware and software companies: Dell went private, Sun Microsystems finally gave up the ghost, and the world finally realized Oracle is predatory and the exodus from them is ongoing.

This is when we founded OC4. To give a little bit of perspective on how different the world was, in 2009 AWS consisted of EBS (at a staggering 100 IOPS), EC2, SQS, and S3. This is in stark contrast to the ~150 (and growing) unique services at the time of this writing. This is the first indication of a concept that will be core to this writing: the pace of innovation is only accelerating. Better get on the train, or you’re going to be left at the station.

And there was this book that we were dropping off at every one of our sales calls called The Big Switch by Nicholas Carr. Ten years later, I can tell you that meets the standard for being a darn good book, as it’s still relevant (and that’s rare for ten year old books about technology).

Early FUD, Early Innovation

In the early days of AWS and The Cloud, it was easy for the legacy hardware manufacturers to take their shots. The FUD strategies consisted of little more than condescendingly asking “what is The Cloud?” and (rightfully) questioning the ability of AWS and other Cloud providers to execute properly on security and on meeting the regulatory requirements needed by various industries.

Experiment, Fail, Learn, Innovate

I remember listening to Dr. Werner Vogels talk about failure at the inaugural AWS re:Invent in 2012. The first re:Invent was tiny - about 5,000 people in attendance. But what Dr. Vogels talked about that day resonated not only with those of us who attended, but in the very way that AWS did business. The simple mantras of “Everything fails all the time,” “Design for failure,” “Embrace failure, it’s a good thing,” and many others were evident in the way AWS was approaching bringing Cloud to the masses. Even though the masses were slow on the uptake. 

In the three years being discussed here, we saw a massive uptick in the number of services from the early days, including core services like Route 53 and VPCs and DevOps Capabilities like CloudFormation and Elastic Beanstalk, and database as a service offerings of DynamoDB and Redshift. But perhaps most importantly, we saw AWS’ constant re-investment in its own infrastructure as well as a lot of under the radar work on certifications and regulatory compliance come to fruition: in May of 2013, AWS received an Agency Authority to Operate (ATO) from the Department of Health and Human Services under the FedRAMP program. All of the sudden, AWS had been vetted by a major branch of the United States Government, and was authorized to perform sensitive workloads for the government.

Pouring Gas on the Fire

In the approximately six years since that time, the pace of innovation as not slowed. To the contrary, it has accelerated. We’ve seen major pushes from Microsoft and Google to compete with AWS for control of the majority of the Cloud space. Alibaba appears to be emerging as a player in at least a regional (APAC) capacity. And because every good story needs a bit of comic relief, Larry and the crew over at Oracle decided to put the word “Cloud” in front of about 1/2 of their product offerings and yell really loud in hopes of keeping their customers (but that’s a separate post for another time).

Meanwhile, Back at the Ranch

While AWS and the other Cloud vendors continued to take massive steps forward, including gaining regulatory approval to run highly sensitive healthcare, financial, and governmental workloads in The Cloud, the majority of Enterprises continued with the old school mentality of “if it ain’t broke, don’t fix it,” and / or “no one ever got fired for bringing in IBM…” 

Enterprises had invested significantly in data center space, and in many cases were still paying down accrued debt on these data centers. Meanwhile, advances in hardware and virtualization made the need for a bespoke data center decreasingly relevant over time. C-level executives who had made big bets on in-house operations were removed from command. And they were removed for one reason: dollars

The CEO and The CFO Get On Board

It’s no secret that a major advantage of Cloud adoption is that we can move Capital Expenditures (CapEx) to Operational Expenditures (OpEx), and in the process, get rid of all of the amortization and tax burden associated with capital assets and the cost of money. 

As we’ve moved from Enterprise toe-dipping to seeing major corporations announce that they are all-in on Cloud, it’s clear that these decisions are impacted not only by the technical side of the C-Level officers, but also by the CEO and the CFO. And while it did take some time for the CapEx to OpEx argument to become clear at this level, many Fortune 2000 companies are now under mandates from their CEOs and their Boards to more aggressively adopt Cloud Strategy.

Timeline: Now

The technical staff has received a mandate from the CEO that Cloud Adoption is a priority. The technical leadership is uncomfortable with this, because change is difficult. The technical leadership is further handicapped by their desire not to lose aspects of their fiefdoms; they do not want to see budget or headcount move to a different group for fear of themselves becoming irrelevant.

In many Enterprises, senior staff take the wait-and-see approach. They claim that they “want to see how it plays out,” and that they “want to avoid vendor lock-in” in their Cloud approach. 

This is where we are today. This is where the gut check comes, and this is where it’s necessary to see some actual leadership among the technical staff.  

Jocko Willink and Leif Babin have a great quote about this concept in the book Extreme Ownership: How U.S. Navy SEALs Lead and Win, when they say:

“Leaders should never be satisfied. They must always strive to improve, and they must build that mind-set into the team. They must face the facts through a realistic, brutally honest assessment of themselves and their team’s performance. Identifying weaknesses, good leaders seek to strengthen them and come up with a plan to overcome challenges. The best teams anywhere, like the SEAL Teams, are constantly looking to improve, add capability, and push the standards higher. It starts with the individual and spreads to each of the team members until this becomes the culture, the new standard. The recognition that there are no bad teams, only bad leaders facilitates Extreme Ownership and enables leaders to build high-performance teams that dominate on any battlefield, literal or figurative.”

Put on your big boy / big girl undies: It’s time to lead.

Bob Dylan Said It Best

The decision regarding which Cloud best fits any given Enterprise’s needs is not one to be entered into lightly. That said, it is a decision that must be entered into. And the one answer that cannot be abided in this decision making process is multi-cloud.

You know why? Here’s what Bob Dylan had to say about it:

Yep - You're Gonna Have to Serve Somebody.

Just like decision have to be made on the Enterprise Level for which Operating Systems will be supported, which networking hardware will be used, which hardware vendors will supply the laptops, and which firewalls will be used, a decision needs to be made regarding what Cloud will be used. And the reasons why are simple:

1. Some great tools exist out there to put a layer of abstraction between your Cloud Provider and your Infrastructure as Code, including the increasingly ubiquitous Terraform from HashiCorp. But these tools can only function at the point of the lowest common denominator. The way that an AWS Lambda behaves and interacts with its environment can be very different from the way a Microsoft Azure Function interacts with its environment. And microservices are just one aspect of the environments that are being deployed. We’ve got storage, networking, managed databases, and a slew of other services to consider. If the choice is made to be Cloud agnostic, then set of tools that will be available in a given Cloud are limited to those tools that are common across all Clouds in use. This alone makes it a bad strategy.
2. Unless an organization is staggeringly large and has completely different Cloud strategies for its different business units, developing a core set of practices and skill sets, as well as developing re-usable templates for security, reporting, and common service will serve the organization significantly better than attempting to implement Cloud generically. You can consider this the “Jack of all trades, master of none” argument.
3.  Enterprises have already made this decision. Think about it: hardware vendors for servers, laptops, and networks. Standardization on software such as Exchange, LDAP, AD, or databases. Even standardization on a lower-level scale pertaining to permitted protocols within the Enterprise. It’s no different with Cloud: in order to maximize the efficiency of a given Cloud, the consumer must adopt an all-in approach.

Bringing It Home

Put simply, there is no one prescription for Cloud adoption in the Enterprise. If there is one thing I learned in my 20+ years as a consultant, it is that every Enterprise has unique requirements and a unique way of approaching their business requirements; it is this more than anything that ends up being the “secret sauce” that differentiates one company from another. As an AWS Expert and a graybeard in general, I’m always going to be skewed towards AWS. 

At the end of the day, the pace of innovation is not going to slow down, and the major Cloud vendors are not going to become more alike - they will differentiate and specialize based on what their customers demand. To this end, the best solution for Enterprise Cloud adoption is to take a tempered approach, perform all possible due diligence to ensure a good fit, and then to go all-in with your chosen vendor.

About the Author

Will Malz is the Founder and CEO of OC4. He holds 8 out of 9 AWS Certification, and serves as a Subject Matter Expert for the AWS Certification process for AWS SysOps Administrator, AWS Solutions Architect - Professional, and AWS Security Specialty. He routinely works in and speaks on Cloud adoption and security best practices as they pertain to highly regulated verticals.