Does AWS Have a Weakness?

The year was 2008, and I was at the Interop IT trade show in Las Vegas, lugging around my analog camcorder, pestering tech vendors for interviews. What’s up, what’s trending? How does your product shape the future?

Funny thing about being a tech journalist at an IT trade show: the vendors are generally terrified of you. Unless they get permission from three layers of management, speaking to a journalist – even to publicize their product – is viewed like walking on glass. A terrible experience, clearly dangerous. Best avoided at all costs.

One lady, I remember, simply cast down her eyes, said “no,” and walked away without a word.

But then I got to this booth for an upstart operation called Amazon Web Services. The young guy manning the booth was fearless. Talk to a journalist, record a video? Sure, he’d love to. No, he didn’t need three layers of approval. After all, why do vendors come to trade shows if not to hawk their wares?

My video from 2008 of him describing AWS is a marvelous museum curio. It’s a look back at a seedling that would grow so massive that it reshaped the entire tech landscape:

AWS, he explained – assuming I hadn’t heard of it – “is part of Amazon.” Modestly, he said that AWS offered “a very small number of core Web services,” for “generally non-mission critical uses in an enterprise environment.”

No one else approached the booth as he spoke, though the exhibit floor was crowded. Why would anyone pay attention to this marginal AWS business?

“Historically, our business has been built on startups who were trying to avoid upfront investment in infrastructure,” he said. He added a quiet claim that would resonate through the years: “We’re able to now offer that to enterprises.”

Fast forward to 2017, and – you know this – AWS stands dominant in the cloud. Its revenues are bounding skyward. The first year that Amazon broke out financials for AWS was 2014, claiming $4.6 billion. In 2016 AWS hauled in $12.2 billion. Deutsche Bank forecasts that the cloud giant will haul in $16 billion in 2017.

It’s now repeated like a mantra: AWS is the leading cloud vendor. It’s dominance, like the array of new tools it releases, seems never ending. Take a deep breath and repeat after me: AWS reigns supreme in the cloud market.

Or does it?

Even at the low end of the forecast, Wikibon predicts AWS’s revenues will jump to $33 billion by 2022; the high end figure is a stunning $48 billion.

The Promiscuous Cloud Market

The history of technology is famously fickle. A high flyer today can make a few wrong turns and soon fall toward oblivion. Amazon’s own Fire Phone debuted in 2014 with rock solid marketing support – and collapsed about 90 days later. Clearly, the company isn’t infallible.

“Perhaps the biggest challenge facing AWS is that of all champions, staying sharp and hungry,” says Bernard Golden, CEO of consultancy Navica. He adds: “All other high-growth technology companies have eventually succumbed to product complexity, internal politics, and poor leadership. To this point, AWS has avoided these issues — the question is whether AWS can continue its growth in the face of the challenges that have afflicted previous high-flyers.”

A key challenge to AWS’s high flyer status: we’re living in a multi-cloud world. Businesses are truly promiscuous in selecting cloud vendors. The Rightscale 2017 State of the Cloud report indicates that cloud users run applications in 4 clouds and are experimenting with 4 more. Some 85 percent of enterprises employ a multi-cloud strategy, up from 82 percent in 2016.

In sum, AWS first mover advantage didn’t translate into monolithic control. If anything, the still wide-open market means that AWS faces challenges from every side.

Amid the churn, Microsoft Azure cloud is growing dramatically. The Rightscale survey reveals that in the enterprise sector – the most lucrative – Azure boosted adoption from 26 percent to 43 percent. Google Cloud, too – remarkably slow to start its enterprise cloud – rose from 9 to 15 percent. (Here’s my analysis of Google Cloud.) AWS gained modestly, from 56 to 59 percent.

We may look back at this period of cloud history and see AWS’s dominance as a short blip, a brief period before a a bevy of large and small competitors eventually catch up and devour market share.

Over the last year, most top public cloud vendors gained far more market share than AWS.

Unenthusiastic about Hybrid

One of Microsoft Azure’s strength is its embrace of the hybrid cloud, a model for which AWS doesn’t exhibit great fondness. Hybrid cloud is clearly the dominant model, used by 85 percent of companies in the Rightscale survey. It’s necessary in some cases for privacy and compliance; many companies need to “hang on” to certain key elements with a private, in-house cloud.

Microsoft “has the broadest solution across the hybrid environment,” says Wikibon analyst Stu Miniman.

Azure is aggressively romancing hybrid cloud users by debuting the Azure Stack appliance, due out this summer. The Azure Stack enables companies to run a version of the Azure Cloud in their own datacenter. Applications are linked between environments, creating a converged private-public union. (Here’s video of my conversation with Microsoft executive Mike Shutz about the Azure Stack.)

AWS lacks a similar product. It is certainly feasible to run a hybrid cloud with AWS and many companies do so. Yet this kind of private-public union seems antithetical to AWS’s approach, which seems to be “public cloud is the cloud.”

AWS’s apparent lack of enthusiasm for the hybrid cloud may be a significant weakness going forward.

Deep Roots are Hard to Move

AWS’s early success was with start-ups and developers. That is, digital natives, not businesses with entrenched infrastructure. Even as it attracted enterprise customers, one of its biggest success stories is Netflix, which is very much a product of the Internet era.

However, “Most organization were not born digital,” says IDC’s Matt Eastwood, SVP, Enterprise Infrastructure and Datacenter. “Instead they have a mix of traditional legacy applications and emerging cloud native applications which frequently need to live side by side on an infrastructure that has the necessary agility, elasticity and security.”

These legacy apps create an opportunity for Microsoft Azure and other cloud vendors with a longer history. For businesses that are heavily invested in traditional workhorses like Microsoft SQL Server, IBM DB2 and Oracle applications, doing a straight “lift and shift” to AWS isn’t feasible.

“If I’m a typical enterprise customer, and I’ve got my hundreds or thousands of applications, it is a longer path to Amazon, there are more challenges and hurdles for me,” says Miniman.

The typical enterprise outlook is: “If I have Microsoft apps, maybe I’ll stay with Microsoft [in the cloud],” he says. “If I’ve got Oracle apps, Oracle has put a huge effort into cloud. And let’s not forget IBM. IBM has lots of customers with lots of apps, and they have lots of ways to get you there.” This is also true of HPE and Dell/EMC: “they have customers who have their environments, and they’re going to offer cloud solutions.”

AWS is well aware of this hurdle and is deploying Amazonian-level resources to address it. In 2016 AWS unveiled its Database Migration Service. The company claims it has performed upwards of 20,000 migrations as of March 2017.

In case AWS’s competitive intent isn’t clear, it offers this helpful article: How to Migrate Your Oracle Database to Amazon Aurora. Yet whether AWS will be able to tempt businesses to its own versions of apps (or otherwise attract legacy-bound companies) remains an open question.

In a similar effort, AWS inked a partnership with VMware, with an eye toward onboarding those companies whose infrastructure is entwined with the default virtualization platform. Currently VMware Cloud on AWS is in technical preview, with an expected roll out later in 2017.

The IoT Jump Ball

In his oft-cited video, Return to the Edge and The End of Cloud Computing (see below), Peter Levine, a general partner with Andreessen Horowitz, opines that as computing moves to the edge, cloud computing will fade. Why? “Because the edge — where self-driving cars and drones are really data centers with wheels or wings — is where it’s at,” Levine explains in his video.

Levine’s viewpoint is debatable, but it does reflect an undeniable truth. Edge computing – the Internet of Things – is a booming market that will reshape consumer and business life over the next decade. Everything from your dog to your coffeepot will have an IoT sensor. IDC forecasts that IoT spending will surge from its 2016 level of $737 billion to a whopping $1.29 trillion in 2020.

As Levine notes, much of the torrent of data from IoT devices will be processed at the edge. However, the data stream will offer too rich a potential to rely on edge processing alone. Undoubtedly, a deep river of data will flow back to central repositories, increasing the need for cloud-based applications. At this point we don’t know the balance between edge and center-based processing.

Whatever the case, IoT lacks a clear leading vendor. In terms of market share, it’s a “jump ball,” with a mess of competing standards and a flurry of vendors trying to build presence.

In this scrum, AWS’s top status in the cloud is less relevant. It is conceivable – especially if you buy Levine’s argument – that IoT could push the Amazon cloud toward the margin.

AWS is clearly focused on the IoT market. In an exceptionally astute competitive move, in late 2016 it unveiled Greengrass, which enables local compute and data caching for IoT. Connected devices will use AWS Lambda, which allows users to run code without managing servers – the so-called serverless compute mode. Greengrass is currently in “Limited Preview” mode.

The Race that No One Has Ever Won

“Nobody can doubt that AWS has been hugely successful, says David Linthicum, SVP at Cloud Technology Partners. “Indeed, I give them credit for making the IaaS public cloud space what it is today.” Still, “there are risks to their continued success, such as market issues, meaning that cloud computing does not grow at the pace everyone is predicting.”

Furthermore, a mixed basket of challenges confronts AWS, Linthicum notes. These include: “the risk of a cross-tenant hack, or an outage lasting more than a few hours. Finally, regulatory issues may limit growth, such as tax advantages for purchasing new hardware and software, or legal limits on what type of applications and data can be placed in the cloud, beyond what is on the books today.”

IDC’s Eastwood notes that, at this point, “Leading AWS deployment challenges stem from performance, SLAs, interoperability and ongoing costs. This will continue as tomorrow’s digitally transformed business looks for an infrastructure that is increasingly optimized for a continuum of application and data services extending from the edge to the core to the cloud.”

That last sentence holds what is likely the greatest competitive challenge for AWS. Maintaining and growing its dominant position will require it to service all areas: cloud, core and edge, which is something that no one company has ever done. Today’s technological landscape, with its byzantine IT complications and chaotic rate of change, has never existed before. Will AWS be the first technology provider to do it all?