Follow Datanami:
January 8, 2019

As Cloud Grows, Is Resistance to AWS Futile?

(Beppe Castro/Shutterstock)

If you’re looking to build a big data application in 2019, it would be foolish not to check out what Amazon Web Services can offer. Whether you’re looking for Hadoop-style processing, graph and time-series databases, or training a neural network, the company has what you need. But is there a potential downside hidden amid this proliferation of easy technological goodness?

Amazon Web Services is by no means a startup, but its financial statements have been resembling one recently. The Seattle, Washington company was initially created in 2002 to host the servers that powered Amazon’s ecommerce business, and in 2006 it opened up its doors up and started providing server, storage, and other IT services for other organizations.

By 2010, AWS had become a $1.5-billion business. Steady if unspectacular growth continued, and by 2015 AWS was a $5 billion business for Amazon, giving a five year-compound annual growth rate (CAGR) of just under 30%. By then, analysts started to take note of the high profitability of AWS, especially compared to its parent company, which poured money into its own expansion instead of recording profits and paying dividends.

By 2017, AWS revenues increased $17.5 billion, a 43% increase over the previous year. The company was bigger than its next 14 competitors, combined, CBS News pointed out. During the company’s AWS re:Invent conference in November 2018, CEO Andy Jassy said the company was at a $27 billion annual run-rate and growing by 46% annually, which would make AWS about the same size as SAP, the fifth largest software company on the planet.

If that growth rate holds up, which is no guarantee, AWS will be a $57-billion business by 2020, and will exceed the $100-billion mark sometime in 2021, equaling the revenues of Microsoft, currently the second-most valuable company in the world. The most valuable company in the world right now is Amazon itself, with a $812 billion market capitalization.

Data Centers Galore

What does a $27-billion data center operation even look like? Nobody really knows. In 2014, when AWS had a fraction of the revenue than it does today, EnterpriseTech editor Timothy Prickett Morgan did the math based on AWS engineer James Hamilton’s re:Invent presentation and estimated that AWS had between 2.8 million and 5.6 million servers across 87 data centers, comprising 28 availability zones.

AWS likely has close to 200 data centers likes this around the world (image source: AWS)

How many data centers does AWS have now? The company doesn’t publish that information, but it does state that it has 60 availability zones, with plans for 12 more availability zones. A recent leak by WikiLeaks indicate that the company had 108 data centers around the world in 2015.

If you assume that an availability zone is composed of 3.1 data centers (which may or may not be accurate), that would mean AWS has more than 180 data centers around the world, with plans for another 40 or so more. Clearly the company is operating at a mammoth scale, fueled by customers’ demand to get out of the business of installing, running, and maintaining servers and other IT gear. But what’s the downside?

Cloud Dominance

In November, Synergy Research Group released its latest estimates on market share for public infrastructure as a service (IaaS) and platform as a service (PaaS) providers. AWS owns 40% of the worldwide public cloud market, according to SRG, while Microsoft and Google together own 25% of the market. (While Microsoft and Google are technically growing faster than AWS, they’re operating on a much smaller base, and have a long way to catch up with AWS.)

Synergy’s Chief Analyst and Research Director John Dinsdale says clouds started to go mainstream in the period from 2014 to 2016, and became “the new normal,” in 2017. By 2018, public cloud started to dominate IT spending in some areas, “sucking up potential growth opportunities for non-cloud technologies and services.”

“Cloud technologies are now generating massive revenues for both cloud service providers and technology vendors,” Dinsdale said, “and our latest forecasts show that while market growth rates will inevitably erode due to the sheer scale of the numbers, the overall market will double in size in under four years.”

Cloud computing will be a $300 billion business by 2020, according to Gartner (By AjayTvm/Shutterstock)

Gartner says the overall cloud market in 2018 was worth $186 billion, and will grow by 62% to become a $302 billion business by 2021. Sid Nag, research director at Gartner, says the growing dominance of the hyperscale cloud vendors creates “both enormous opportunities and challenges for end users and other market participants.”

“While it enables efficiencies and cost benefits,” Nag says, “organizations need to be cautious about IaaS providers potentially gaining unchecked influence over customers and the market. In response to multi-cloud adoption trends, organizations will increasingly demand a simpler way to move workloads, applications, and data across cloud providers’ IaaS offerings without penalties.”

Too Big To Fail?

Some commentators have begun speculating that Amazon and AWS have become “too big to fail,” the label that has been applied to big Wall Street banks during the financial crises of 2008, which necessitated a bailout by the Federal Government measured in the trillions of dollars.

During a 2016 IT industry event hosted by Canalys, a market analyst firm, Canalys founder Steve Brazier indicated that AWS was going down the road of becoming too big to fail. “If one of these companies goes down, hundreds of thousands of other companies go down too,” Brazier said, according to this story in The Register.

Amazon CEO Jeff Bezos disagrees. “Amazon is not too big to fail,” Bezos said last year during a company call, according to this story in Forbes. “In fact, I predict one day Amazon will fail. Amazon will go bankrupt. If you look at large companies, their lifespans tend to be 30-plus years, not a hundred-plus years.”

Clearly, AWS is not going to go bankrupt any time soon. For one, it’s just making way too much money. But it also appears that AWS is pretty darn good at its job.

While nobody has forgotten the massive AWS outage that occurred in March 2017 due to human error, we also have yet to hear about any the big cloud providers suffering the types of major security breaches that have impacted big companies like Equifax, Sony, and Marriott that have opted to keep their servers on premise.

Competitive Threat

However, some analysts have begun wondering whether AWS has gotten so big that it’s beginning to stifle innovation in the market. Some vendors are also taking action in response to the threat.

For example, in response to AWS’ success in turning open source technology into profitable businesses, MongoDB, Redis Labs, and Confluent decided in late 2018 to impose additional license requirements on companies that would run one or more parts of their offering in the cloud, with the end result being that their open source software is a bit less open.

The merger of Cloudera and Hortonworks was also driven in some part by the success that AWS has had creating data lakes on its Simple Storage Service (S3) object store and Hadoop-style big data processing via the Elastic MapReduce (EMR) service. Jassy says customers are running 10,000 data lakes on AWS, which dwarfs the number of joint customers (2,500) shared by Cloudera and Hortonworks.

Forrester senior analyst Paul Miller doesn’t see a major threat from AWS that would stifle innovation, at least not yet.

“I’m certainly not seeing any evidence of a regulator concern at the moment, particularly because of the speed with which the competition is growing,” Miller tells Datanami. “AWS is the largest, by far, but Microsoft and Google are growing faster than AWS – granted, from a lower base, but they’re growing faster. Therefore, it would be difficult to make an argument to a regulator that said AWS is monopolistic. They’re not behaving in that fashion at the moment.”

That’s not to say that AWS customers should not be concerned by the potential for lock-in. All vendors desire their solutions to be “sticky,” in the sense that customers don’t abandon them, and AWS is no different. Stay tuned to Datanami for Miller’s insights on how to utilize AWS offerings without getting stuck.

Related Items:

Cloud Backlash Grows as Open Source Gets Less Open

AWS To Build You a Data Lake in ‘A Few Clicks’

A Rare Peek Into The Massive Scale of AWS (EnterpriseTech)