If by 2016 it was clear that big chunks of enterprise data center workloads would soon be running in the cloud, 2017 was the year enterprise cloud migration shifted into high gear. This migration reverberated throughout the data center industry, bringing about big changes for technology vendors and data center providers.
Data center infrastructure vendors – the likes of Vertiv (formerly Emerson Network Power) – have realigned their strategies, resources, and product portfolios to cater to clients who operate data centers that host cloud infrastructure. IT hardware vendors continue to struggle with deterioration of their core value proposition driven by cloud providers’ development of their own hardware design and supply chain capabilities. As an example, Hewlett Packard Enterprise, one of the leading vendors in this category, this year decided to get out of the business of selling commodity boxes to cloud giants altogether.
The biggest data center providers have seen unprecedented growth rates driven primarily by cloud leases. As more and more workloads move to the cloud, cloud providers need more and more space and power to support those workloads, and landlords like Digital Realty Trust and CyrusOne have benefited from this trend handsomely. To get more cloud business, data center providers have had to bulk up, driving a wave of consolidation that has left few independent players of significant scale in the US while also steadily consuming companies in Europe and Asia-Pacific.
Here’s a list of the year’s biggest data center provider acquisitions and our coverage of the deals (there were many more -- these are just the biggest ones):
- Why Iron Mountain’s IO Deal Spooked Wall Street
- Shaw Sells Data Center Provider ViaWest to Peak 10 for $1.7B
- DuPont Fabros Deal Gives Digital Realty More Firepower in the War for Cloud Deals
- Cyxtera Puts a Fresh Spin on CenturyLink’s Former Data Center Empire
- Digital Bridge Buys Vantage, Silicon Valley’s Largest Wholesale Data Center Firm
(While Equinix’s blockbuster acquisition of the massive Verizon data center portfolio was closed this year, it was announced in December 2016 and as a result didn’t didn’t make this list.)
As the cloud business grows, all the biggest cloud providers are now conceding that many of the large enterprises they want to serve have no intention of moving all their applications to the cloud. For a variety of reasons, lots of computing infrastructure will remain in on-premises data centers or colocation facilities -- both core and edge -- and this fact really came into focus in 2017. The year saw hybrid strategies by all major cloud providers take shape – each of them a different combination of own technology and partnerships with new and traditional enterprise vendors.
The Great Migration
Every senior enterprise data center and infrastructure operations staff member I spoke to privately at Gartner’s IT and operations conference in Las Vegas earlier this month was either in the middle of a major cloud migration project or actively planning one. Most of them were working for household-name companies.
Contract provisions often prohibit cloud providers and customers from disclosing their business relationships publicly. But some deals are public. In the fourth quarter alone, AWS announced major cloud services agreements with GE, The Walt Disney Company, National Football League, Turner, Toyota, and FICO. Microsoft’s recent enterprise cloud wins include GE as well, but also United Technologies, Bank of America, Halliburton, and UBS. Not exactly “born-in-the-cloud” startups.
The great enterprise cloud migration is evident in revenue growth rates reported by the cloud giants. Total revenue in cloud infrastructure services – now a $12 billion market -- is growing 40 percent per year, according to the analysis of cloud providers’ third-quarter earnings by Synergy Research Group. (The analyst firm includes IaaS, PaaS, and hosted private cloud in its cloud infrastructure services category.) Already lightyears ahead of the others in terms of market share, AWS continues to grow its piece of the pie. But its biggest competitors are reporting healthy growth of their own. In fact, Microsoft, Google, and Alibaba are growing their cloud revenues faster than AWS is, according to Structure.
Adjusting to the Fact of Hybrid Cloud
As they compete for enterprise cloud dollars, public cloud providers are figuring out ways to enable those customers to connect their on-premises environments and cloud services. These hybrid cloud strategies were a huge part of the conversation this year.
AWS and VMware started rolling out the first products and services born out of a partnership the two companies announced in 2016. They started by enabling customers to spin up VMware servers in select AWS availability regions and use AWS for backup in place of their own disaster-recovery data centers. Microsoft started shipping Azure Stack, the on-premises Azure environment that can be integrated with its public cloud (but doesn’t have to be). Google Cloud announced a series of partnerships around hybrid cloud, including with hyperconverged infrastructure leaders Nutanix and Scale Computing. Google also partnered with Cisco -- a deal that's especially interesting, aiming to combine Cisco’s HyperFlex hardware stack with a Google Cloud software stack (including the massively popular container orchestration platform Kubernetes) for on-prem deployments.
Cloud is Extremely Resilient but Not Infallible
Many of the companies already running important applications in the cloud learned early this year what should be a valuable lesson for everybody handing their workloads to a cloud provider: while your apps may be running in the cloud, ensuring uptime is still up to you.
The AWS outage in March, caused by a single mistyped command by an engineer, brought down popular services like Expedia, Coursera, Medium, Quora, and Slack (among many others) and, according to one cyber-risk modeling company’s estimate, cost S&P 500 companies about $150 million in total losses.
According to experts we interviewed following the massive cloud meltdown, AWS customers could withstand the outage -- and many of them did -- by investing extra time and money in architecting their applications in a more resilient way. Essentially, it comes down to creating redundancy across multiple availability regions, redundancy across multiple cloud providers, and using caching services that store multiple copies of your files across many different locations.
To be sure, hyper-scale cloud platforms are incredibly resilient (more resilient than many custom in-house enterprise platforms), and companies that operate them hire some of the brightest minds in the business to ensure they stay that way. But these systems are not infallible. As more and more workloads shift to the cloud, enterprises share the responsibility for their application uptime with their cloud providers.
See You Next Year!
With that, the staff at Data Center Knowledge would like to thank you for visiting our site this year. We hope we’ve been useful to you and had a positive impact in your professional life. We are taking a week off to refresh and reflect, but we’ll be back on January 2nd. We have lots of new ideas for making our site even more useful to you in 2018, and we cannot wait to start bringing them to life in the New Year.