IaaS in 2016: Who can challenge Amazon?
Infrastructure as a service (IaaS) is a maturing segment of the cloud computing market. In its 2015 Cloud Computing Hype Cycle, Gartner placed IaaS on the 'Slope of Enlightenment', which is the last stage before mainstream adoption (a.k.a the Plateau of Productivity).
Also approaching the SoE is 'IaaS+', which provides additional services on top of basic compute, storage, and networking (such as database services and content delivery networks) but falls short of those offered by full-blown PaaS (Platform as a Service) providers. Meanwhile, 'Private IaaS' (single-tenant cloud infrastructure, hosted by a provider or on the client's premises) currently languishes in the Trough of Disillusionment.
The dominant IaaS player in the public cloud is Amazon Web Services (AWS), which launched in 2006 with EC2 (compute) and S3 (storage) services. In Amazon's most recent quarterly results (Q2 2016, ended 30 June), AWS generated $2.9 billion in revenue, representing 58 percent year-on-year growth:
AWS's $2.9 billion is just 9.5 percent of Amazon's total $30.4 billion revenue for Q2 2016 -- but it's a profitable segment, generating $718 million in operating income for the company.
AWS's market dominance in cloud computing is based on its long track record and extensive feature set, which has given some companies the confidence to abandon traditional IT infrastructure altogether. In his keynote at the 2015 re:Invent conference last October, Andy Jassy, then senior VP (now CEO) in charge of AWS, highlighted the number and breadth of organisations that are going "all-in" on the company's cloud services. These pioneering organisations -- among them are well-known companies like Netflix, Intuit, and Juniper -- are decommissioning their private datacentres and running their businesses entirely on AWS's on-demand compute, storage, networking, and other services. Until recently, such a strategy would have been regarded as madness by most CTOs.
Amazon Web Services CEO Andy Jassy, onstage at the 2015 re:Invent conference. (Image: Amazon)
Amazon may be the dominant public cloud infrastructure player, but there are plenty of would-be challengers, including Microsoft's Azure and Google's Cloud Platform.
In its recent 2016 Magic Quadrant for Cloud Infrastructure as a Service, Gartner placed Microsoft along with Amazon in the 'Leaders' quadrant, with Google on its own in 'Visionaries' and the remaining seven providers -- including big names like Rackspace, IBM (SoftLayer), and VMware -- lumped under 'Niche Players'.
According to Gartner, AWS supports a broad range of use cases (including enterprise and mission-critical applications), has attracted an ecosystem of open-source tools and a network of partners, and is a mature yet agile cloud provider that's the "safe choice" for customers seeking rich functionality and a long-term market relationship. On the downside, Gartner notes that customers will require expertise (either in-house or via professional services) to get the most from AWS's multiple offerings, and that its granular price structure is complex.
Microsoft, which entered the IaaS market in 2013 with Azure Infrastructure Services, is praised by Gartner for its good integration of IaaS and PaaS (Platform as a Service) components that seamlessly extend and interoperate with on-premises Microsoft infrastructure. Azure is less feature-rich and less mature than AWS, but it's growing fast and its pricing is comparable. However, Gartner notes that some enterprise-class functionality is yet to be fully implemented, documented, and supported, and that third-party Azure experts are currently limited in number.
Google also entered the IaaS market in 2013, with Google Compute Engine. Gartner sees Google's ability to run technology platforms at scale -- honed on its internal operations -- as a key advantage, along with a comprehensive vision for, and experience with, the lifecycle management of cloud-native applications. Big data, analytics, machine learning, and batch computing are also key strengths. Where Google lags AWS, Microsoft is in the breadth of its feature set, which is missing some capabilities both for large organisations and startups, says Gartner. Google is also still learning to engage with enterprise and mid-market customers, according to the analyst firm.
How big a pie are these and other cloud providers competing for?
Gartner estimates that worldwide market for public cloud services will be worth $204 billion in 2016 -- a year-on-year growth rate of 16.5 percent. Although at $22.4 billion, IaaS represents only 11 percent of the cloud computing total, and it is the fastest-growing segment, with a growth rate of 38.4 percent.
In a statement, Gartner analyst Sid Nag noted that "IaaS continues to be the strongest-growing segment as enterprises move away from data center build-outs and move their infrastructure needs to the public cloud... Certain market leaders have built a significant lead in this segment, so providers should focus on creating differentiation for success."
RightScale's State of the Cloud surveys
To get a picture of how the cloud computing industry has developed in recent years, it's worth examining the annual State of the Cloud surveys -- mostly canvassing IT pros in medium-sized to large tech companies in North America -- conducted by cloud portfolio management company RightScale. (For more on RightScale's State of the Cloud surveys, see this article on ZDNet's sister site Tech Pro Research.)
When asked about the perceived benefits of moving workloads to the cloud, the number-one advantage, cited by 62 percent of respondents in the 2016 survey, was 'Faster access to infrastructure'. This was also the largest increase over the 2015 survey, up from 57 percent. Next came 'Greater scalability' (58 percent), 'Higher availability' (52 percent), and 'Faster time to market' (52 percent -- the second largest increase from 2015, up from 48 percent).
Interestingly, cost-related issues -- 'Move capex to opex', 'IT staff efficiency', and 'Cost savings' -- propped up the 2016 benefits table, all being cited by less than 40 percent of respondents. This finding supports the first of Gartner's Top 10 Cloud Myths -- namely that "Cloud Is Always About Money".
Security concerns are often given as a reason to be wary of the cloud, and indeed were the No. 1 challenge reported in RightScale's 2014 and 2015 surveys. In 2016, however, security was overtaken by 'Lack of resources/expertise', which moved from joint last to first place in the 'Cloud challenges' list between 2014 and 2016. Again, this supports a Gartner Cloud Myth (#6) -- that "Cloud is less secure than on-premises capabilities".
Other fast-rising cloud challenges reported by RightScale are 'Managing multiple cloud services' and 'Managing costs' -- noticeably operational rather than early-stage adoption headaches.
A lot of cloud computing services are bought to support 'agile' IT -- to develop and run new cloud-native apps and trial various digital transformation projects. This trend is evident in the rise of DevOps and containerisation in RightScale's surveys: DevOps adoption rose rapidly, from 54 percent in 2013 to 74 percent in 2016, and it's more prevalent in enterprises (81 percent adoption in 2016 versus 70 percent for small- to mid-sized businesses). However, only 21 percent of enterprises have adopted DevOps company-wide: 31 percent adopt it at the business unit/division level, while 29 percent restrict it to projects or teams.
As far as DevOps tools are concerned, Chef and Puppet have ruled the roost to date, with Docker, in particular, hard on their heels (in the 2016 survey, an additional 35 percent of respondents 'plan to use' Docker, compared to 19 percent and 18 percent for Chef and Puppet, respectively). Also beginning to make their presence felt in RightScale's recent surveys are container orchestration tools such as Kubernetes, Swarm, and Mesosphere.
When it comes to public cloud providers, it's no surprise to find AWS way out front in RightScale's surveys, with 57 percent of respondents running some applications on AWS in 2016.
However, Microsoft's Azure is making headway, showing the most growth between 2014 and 2016. The most recent survey also shows that a higher percentage of organisations are experimenting with or planning to use Azure services.
The impression that Azure is gaining on AWS is strengthened by some recent Morgan Stanley research, reported by GeekWire, which found that 31 percent of the survey sample expected to be using Azure IaaS by 2019 versus 30 percent planning to be using AWS. The respective 2016 figures were 12 percent and 21 percent:
Before reading too much into this, it's worth noting the small survey population -- just 100 CIOs (75 in the US, 25 in Europe).
Bimodal IT & cloud migration
We noted above that 'agile' IT projects tend to take a DevOps approach, using an array of cloud-native techniques such as containers and orchestration tools. Such innovative 'digital transformation' projects drive a lot of IaaS business, but -- as AWS CEO Andy Jassy pointed out in his 2015 re:Invent keynote -- organisations are increasingly moving to the next stage, which is "all in" on cloud infrastructure. This means moving core business systems to the cloud, and not just innovative or exploratory projects.
Gartner calls this a 'bimodal' approach to the cloud and documents a range of characteristics for each mode:
Of course, the real world is messier than this simplified binary scheme, and as Gartner analyst David Mitchell Smith puts it: "At the end of the day, you have to look at each workload separately."
In the 'Agile' world, the latest development is 'serverless computing' -- which doesn't mean that servers are no longer involved, but rather that DevOps teams no longer need to worry about them. The idea is that 'events' of some kind -- data flowing from IoT devices is a common example -- can be set up to trigger a response from the cloud provider, which automatically provisions the necessary infrastructure for as long as it's required. In effect, DevOps teams are outsourcing the 'Ops' function to the cloud provider.
How AWS Lambda works.
Amazon kicked off this market, announcing its AWS Lambda service in November 2014 and gaining a head-start over the competition. Google Cloud Functions appeared in February 2016 and remains in 'alpha', while Microsoft's Azure Functions -- which can be deployed on-premises as well as in the public cloud -- was announced at Build 2016 in March and is also currently in preview mode.
Outlook
Amazon (AWS) still dominates the cloud infrastructure market and is probably the 'go-to' option for many organisations seeking a public cloud solution. However, Azure is gaining momentum and will appeal particularly to businesses with existing on-premises datacentres built on Microsoft's server software stack. Google has some catching up to do against the two market leaders, but is clearly the 'one to watch'.
Going forward, we can expect plenty of hype and vendor dynamics centred around containers and container orchestration, serverless computing and the various sub-flavours of cloud infrastructure -- IaaS+ and private IaaS.
If you're confused about which cloud is the best home for your workloads, RightScale's free-to-use Cloud Comparison tool, which is updated quarterly, is a good place to start.
Read more about cloud computing
- Stop saying the cloud is just someone else's computer - because it's not
- Where AWS is headed: Every function as a managed cloud service
- AWS Lambda garners interest, production workloads as serverless world evolves
- Microsoft releases preview of new Azure 'serverless compute' service to take on AWS Lambda
- Infrastructure as a service cloud computing revenue to surge by 2020