The momentum in enterprise adoption of cloud technologies is undeniable. The tipping point of cloud adoption is here.
Enterprises are no longer asking if they should utilize cloud technology. They have done so or are rapidly making plans to do so.
In-house and hybrid cloud solutions were the logical first step in enterprise cloud services. It made sense in the early years of cloud technology to keep some business data and functions internal and gradually engage with external cloud services as the business expanded, older technology became obsolete and the organization’s workforce embraced mobile.
The world of technology, however, moves at its own light-speed pace, and in a few short years, these “toe in the water” approaches are already non-optimal, inefficient and inflexible. Thus, enterprises who were early adopters or newcomers to the cloud are now asking “Should we go all-in with a single cloud provider, such as Amazon Web Services”?
In order to answer that question we need to think of a cloud provider as a technology and platform choice and not a vendor choice.
Back when ACME Decided in Favor of “All-In” with Oracle
A fundamental tenet of risk management is not putting all your eggs in one basket. That principle is one of the primary motivations for enterprises to continue with multi-cloud strategies. Doing so potentially spreads the risk of service failures over more than one provider, but does that always make sense in the final cost-benefit analysis?
Five years ago, the CIO at ACME, Inc. pondered whether the company should go all-in on their database strategy by contracting with a single provider, namely Oracle. Considering the enterprise-wide commitment involved, it was a fair question. There would be ongoing investment over 5 to 7 years. Implementation would consume a year. It would be two years before real benefits were felt and – if it all went south – another year to back out.
When seeking board approval, surely there were questions regarding so many eggs in one basket. However, there were many solid reasons on the plus side:
- Their technology and feature sets were mature as was their relationship with middleware and support providers;
- They are a solid provider and a healthy company;
- Their product was widely implemented even within competitors’ organizations.
Lastly, they were the only company with the technology, experience, resources and support required to match the scale of implementation that ACME needed.
So, ACME decided to go all-in with an Oracle technology decision and never looked back.
Comparison to All-In with AWS
Even though the products and business model of companies such as Oracle, SAP or IBM differ significantly in some aspects compared to Amazon Web Services, a line of thinking similar to that which led ACME to choose a single technology provider is applicable:
- AWS offers a comprehensive technology platform, broader in comparison to even Oracle in the above example.
- In terms of feature sets, services and performance of its SaaS and IaaS offerings, no one comes close to their coverage.
- AWS’s vast infrastructure of compute and data resources is scalable to match small SaaS providers up to global organizations such as Nasdaq, Johnson & Johnson and the CIA.
- The AWS ecosystem of partners, developers and support providers is the largest among competitors and continues to accelerate.
Given the breadth and depth they offer in services, resources and expertise, there should be no reservations with regard to AWS’s platform meeting the cloud computing needs of small and large enterprises now and for years ahead.
The Arguments against a Single Cloud Provider
Thus, one rationale for maintaining or implementing a new multi-cloud strategy seems to have been neutralized. What other concerns are there?
Spreading Projects among Clouds
It is tempting to think that your enterprise could migrate internal or external services to multiple cloud service providers. You may have existing relationships or development projects with Google, for instance that you would like to maintain. Plus your organization could be planning to co-locate compute resources using OpenStack in addition to availing themselves of Azure. While you are at it, why not throw AWS into the mix as well.
As far as cloud computing has come in the last decade, mixing and matching cloud solutions in this manner is still not practically feasible, especially on a large scale. If your business can afford to silo data, apps and services, it might work, but then you would be quickly falling behind competitors who are adopting more integrated solutions.
The truth is that Cloud-to-Cloud integration or true portability, scaling and workload sharing solutions are still in their infancy. Docker is the best thing going in this regard with its ability to containerize apps to run on any OS, but it is still very much an early adopter technology.
We Want to be Portable
Our clients frequently put forth a requirement that they maintain complete control over their cloud solutions in order to be able to pick up and move to a different provider. That means choosing solutions that are easier to migrate. They are trading off flexibility for portability. Versus adopting a single platform, portability requires additional implementation and ongoing maintenance.
If you decide later that you want to move to another provider, it is always going to be a migration effort no matter what. And migration may not be as smooth as you had hoped. Furthermore, to achieve portability means having to forgo taking advantage of time-saving products offered on one platform that are not on another.
For example, Amazon Elastic Load Balancer, Amazon RDS for database, or Amazon Redshift for data warehousing save huge amounts of time and increase agility. These are solutions not offered by competitors’ platforms. In our experience, we find customers start off with a hard requirement of portability, but as they learn more about additional features of AWS and the economic value of their unique cloud services, portability moves down the list of priorities.
It seems a logical tactic to play one cloud provider off another in order to procure the best deal. You may gain leverage if your choice is between Google and Azure, but Amazon is not going to budge on price. Amazon publishes their pricing online for everyone to see. To add, with competitors you risk a price rise later, whereas AWS has dropped their prices 49 times already.
Other providers may increase prices, change their services or stop innovating. In the case of Amazon, that would be like a river reversing course and heading back into the jungle. Lowering prices, innovation and continuous improvement are part of AWS’s DNA. It is their flywheel of success. The exact strategy of other providers is murkier.
If Amazon, Microsoft or Google disappear anytime soon, then it would probably be the result of drastic, worldwide changes in priorities due to a lack of food and water to sustain us. Other than a calamity of that scale, the world is still going to be focused on growing the Internet and keeping personal and business apps running.
AWS is essentially the backbone of the Internet today with nearly 40% of traffic running through its infrastructure. If Amazon goes down, the world’s Internet goes down. In such a drastic scenario, many other cloud providers would certainly not be immune to failure.
Is AWS Too Big to Fail?
But, what if the worst were to happen and Amazon saw bankruptcy on the horizon?
Before that could happen, they would have alternatives of which to avail themselves, including a buyout by a well-capitalized company such as Microsoft or Apple who would instantly acquire title to being the number one cloud provider on the planet.
In any case, a complete failure of AWS would not be sudden. It would likely take years if it happened at all.
Even if your organization went with a multi-cloud strategy, there is no guarantee that one or more providers could maintain their SLA and uptime obligations. One way to measure that likelihood is to study the health of the provider in terms of number of incidents per service. Not all providers let clients see this data, AWS on the contrary is very public about it. A healthy provider shows a trend of fewer critical issues per service over time, and AWS exhibits robust health.
Reliability and uptime, however, are not the sole responsibility of the cloud provider. Any enterprise needs to design their services architecture to remove as many single points of failure as possible and to have failure recovery plans in place. We call this design for failure principals. In that regard, you can design and control your own SLA.
The Benefits of Going All-In
Focus. Life gets simpler when you can focus on a single platform for the long term. Less time is spent on maintenance and more is spent on bringing value to your customers and the business. Finding employees within a narrower field of expertise is easier. You have enough on your plate with the myriad technologies and development platforms you already support.
Adding Value at the Right Level. Managed cloud services allow your enterprise to concentrate effort on improving products and making customers happy. Less time should be spent on the heavy lifting of developing proprietary solutions for managing compute and data resources.
For example, AWS Redshift already provides data warehousing that grows according to the amount of data you are processing. AWS’s Kinesis service and their client library allow your business to tap into real-time, diverse, distributed data streams and set up sophisticated dashboards to monitor and filter this data. With the advent of the Internet of Things, this is going to be an essential capability that would be difficult to develop and manage in-house. Besides, with AWS you can count on a scalable solution and only paying for what you use.
Clearly, we try to educate organizations away from a multi-cloud strategy, especially if this is their first foray into the cloud.
The decision is not about choosing a service provider, but one of choosing the correct technology platform.
It is not about who has the most services for the best price at the highest level of reliability. It is about who has the best vision, the best track record, the highest rate of innovation and will still be the leader in providing cloud services when the next 3-year technology strategy discussions begin again.
Going all-in with AWS is our recommendation, but not just because they are far and away the global leader in cloud services. That position is a natural side effect of their vision, their ability to execute and their strong pace of innovation going forward. All of those qualities together bring tremendous benefits to your organization while reducing your risk.