[This article was written by Udi Nachmany.]
If we look at market growth rates and the concentration of power within a handful of providers, using the terminology coined by Geoffrey Moore, it could easily be argued that the public IaaS cloud market is in the tornado and approaching Main Street. Despite a lack of publicly available market share data, for many in the industry, it seems like a two or three-player race, with one clear “gorilla”: Amazon Web Services (AWS).
In this article I will attempt to briefly characterize the competitive positioning of the key players in the public IaaS market, and highlight some of the alternative strategies used by other providers to carve their own niches. The question that needs to be kept in mind is, can anyone else survive in the face of the steep competition presented by the two or three American mega-clouds?
Guesstimating the Numbers
So just how big are the big players? With a lack of publicly-available numbers, we have to make do with what the main players are saying about themselves. Amazon just recently — for the first time — reported its financials for AWS separately from the other business unit, at $1.75 billion in revenues and a 17% profit rate. Amazon execs have also gone on to claim that Amazon has five times the compute power, and four times the revenues, of all other major players combined. These are useful benchmarks, although different companies measure different things: Microsoft combines its public-cloud and public/private cloud revenues from Azure, Office 365 and Dynamics CRM Online into what it calls “Commercial Cloud”. Nevertheless, they suggest a like-for-like IaaS market share for AWS of around 80%. This shouldn’t surprise people who have been familiar with the market since its early days, as Amazon was early to market with what turned out to be a winning bowling-alley approach.
In the early days of the market, Microsoft and Google took the PaaS route with their Azure and Google App Engine offerings, respectively. They offered their expertise to allow user-level simplicity: drop your code on our PaaS and we’ll take care of the rest. Amazon, on the other hand, took another route, providing users with virtual machines for them to configure, orchestrate and use as they saw fit. Despite the power of the Microsoft and Google platforms, developers connected more with Amazon’s approach, which gave them the power to experiment with what was still a nascent computing model, commercially. Fast forward to now, when AWS manages to cement its dominance through an extremely rich offering — claiming to have launched almost two services per day in 2014 — and the sort of customer-oriented commercials we come to expect from Amazon.
Catching Up with Amazon – the Mega-Cloud Angle
Google’s and Microsoft’s attempts at catching up with Amazon has forced these companies to really sharpen their respective value propositions, and radically change some of their long-held precepts.
Arguably, the most impressive transformation has been at Microsoft. Formerly derided as The Empire by many an open source developer, under its current CEO, the company is working hard to reinvent itself as a friendly, almost open platform. Through actions such as open sourcing ‘.net’, offering support for selected Linux distros, or radically changing their sales compensation to match the new reality of the cloud, Azure has managed to move away from being perceived as “the Windows cloud”. All this may be just phase one of a plan which would now build on the company’s unparalleled — certainly where internet-age companies like Amazon or Google are concerned — access to the traditional IT channel. Whether Azure is great or just ‘good enough’ is besides the point when you’ve spent 30 years cementing direct and indirect revenue channels into almost every enterprise in the world.
Google, on the other hand, believes that if you build it (and by it, I mean a great cloud platform), they will come. One of my favourite Google-oriented quotes is by Hadoop co-creator Doug Cutting,
Google is living a few years in the future and sending the rest of us messages.
Google likes to claim that it employs the world’s smartest people, and it is working on arguably the coolest new innovations like self-driving cars, hot air balloons with internet, and the like. Google Cloud Platform’s (GCP) positioning seems to be what it is for the purpose of driving its own innovations, Google has built the best cloud possible, and is now sharing the same infrastructure with its users. Given the sheer size of AWS and Azure, it remains to be seen if that approach can win against the former’s first-mover-advantage and the latter’s hold on the channel. Being more focused than most around container deployments is also a potentially interesting angle for Google.
Is There a Market Beyond the Top Companies?
In light of this question, let’s explore five vectors being pushed by providers worldwide:
- Geography. There are local players in almost every country, from China to Canada to Switzerland to Israel — many of whom might not be with us in a few years. Being local would be tactically useful when serving the public sector in many countries, and strategically useful if the local private and public sectors are worried about data privacy in the post-Snowden age.
- China. The world’s largest economy seems to be dominated by two cloud clusters: the first is comprised of government-owned providers, while the other includes private sector enterprises like Alibaba and Tencent. It is an interesting competitive environment locally for sure, but even more fascinating when we consider that while publicly-traded giants such as Alibaba want to win in China, they are also keen on taking the battle to Amazon and Microsoft.
- Product/market. We’ve recently seen the launch of ARM-based public clouds, while others are focusing on containers, burst capacity/extreme elasticity, on-metal machines and other departures from the virtualization theme. Other players such as Digitalocean are targeting under-addressed segments with uniquely relevant — and compellingly simple — offerings.
- Technology/architecture. The tectonic plates are shifting between the traditional VMware-based clouds — viewed by some as the safe albeit limited choice — and newer frameworks such as OpenStack. This open alternative is rapidly gaining acceptance as a production-viable alternative, with the likes of Wal-Mart deploying 100,000 OpenStack nodes. While the openness and flexibility of OpenStack explains its popularity within the developer community, the possibility of using a similar framework across the private and public clouds is also driving enterprise adoption.
- Hybrid cloud. According to recent press coverage, companies such as VMware, IBM and HP are looking to use their private cloud capabilities as a lock-in mechanism. While possessing technological chops and an enterprise customer base, they are perhaps being realistic about their chances of beating AWS at its own game. Their bet seems to be that with full-on enterprise adoption, the hybrid cloud will emerge as the winning model. In this reality, a CIO would prefer to work with one managed cloud vendor across clouds if they can.
As I have tried to demonstrate, there is still a lot going on in the public cloud market in terms of competitive potential. And like in a myriad of other competitive scenarios which we have lived through, compelling value propositions will be key in creating market space even in the tornado. Lastly, while the big players enjoy dominance, as some have pointed out, they also have to be wary of not attracting unwanted attention from the regulator for what is ultimately a commodity. So the potential success of smaller players in creating their niches, is good for those emerging players, in a way good for the major players, and ultimately great for the customer.
Read more about that: Amazon, Microsoft and Google: The Cloud Leading Trio