It has been a tough last few years for IT Infrastructure and Operations (I&O) leaders aiming to optimize their organizations’ cloud computing. They have gotten experienced enough with cloud computing to know what they should deliver to the business in hybrid cloud (computing capability that spans public clouds and private clouds), yet they have been unable to deliver it. This has left them with segmented public and private cloud efforts: They influence public cloud use while building private cloud capability, expecting to later achieve hybrid cloud capability providing unified control of connectivity, resource utilization / spend, security / compliance and proprietary vendor service dependencies taken (dev/test services, database services, etc.).
The longer they go without delivering hybrid capability, the further their application owners are empowered to implement their own views of optimal cloud usage—often across an array of public and private cloud vendors. In the large majority of enterprises using multiple clouds to date, this has been done with local optimization in pockets of application ownership typically within separate business units. Even for the minority of enterprises that have progressed to a high level of cloud maturity (where business needs are surfaced and planned in cycles for global optimization), lacking hybrid capabilities that automate policy enforcement across clouds makes it nearly impossible to avoid bespoke application implementations. The longer the snowflakes pile up from variance in security approach, capacity commitments, proprietary services, etc., the greater the future operational pain at scale and difficulty migrating applications across cloud infrastructures for cost optimization.
As enterprise infrastructure spend shifts substantially to cloud, it won’t be long before CEOs and CFOs are asking CIOs to have their I&O leaders drive spend optimization, expecting typical approaches to supply chain management will apply, for example in consolidating to fewer vendors to maximize discounts and reduce training and support costs. These business leaders are likely to be surprised by their lack of flexibility given steep cost estimates to migrate applications across clouds that show the degree to which control of their spend has been ceded to the array of vendors. The stakes for I&O leaders in getting to hybrid cloud capability are clear. The past few weeks of industry spring events have offered evidence that 2017 will finally be a turning point in enabling them with a path to start to get a handle on this dynamic.
First, at the Open Networking User Group Spring (ONUG) 2017 conference, a group of over 100 enterprise I&O executives across Financial (ex: Bank of America, Citigroup), Retail (ex: eBay, Gap), Technology (ex: Salesforce, Tesla), etc. industries showed progress in moving from their compiled enterprise user group view of hybrid cloud capability needs to outcomes from an engaged vendor ecosystem. Most specifically, working with vendors like Cisco Systems and Huawei in the SD-WAN space, they have progressed to an API and architecture for an Open SD-WAN Exchange. They previously identified this capability as a top-five need for improving adoption and use of hybrid cloud, because it will reduce a lot of complexity in use of SD-WAN across providers and across physical company locations needing consistent performing connectivity to public and private clouds. They are working similar efforts across their other identified needs, covering unified / hybrid orchestration, monitoring with analytics and security services by pushing for vendor responses with the purchasing weight of the $100B+ per year in IT spend they represent. Some level of these capabilities, particularly unified orchestration, has been available from vendors but has not been widely adopted out of lack of a software offer model for private deployment (since most have been SaaS) as well as concerns over their architecture and readiness for enterprise scale.
Of the public cloud vendors, Microsoft provided the strongest hybrid vision at the ONUG event with its Azure Stack private cloud platform delivering management and orchestration capabilities enabling application portability across Azure public cloud and Azure Stack private clouds. Two weeks later at its Build 2017 event, Microsoft followed up with initial enterprise case study examples from Ernst & Young and Carnival Cruise Line of hybrid deployments of Azure Stack on-premise used in conjunction with public cloud Azure. Ernst & Young is using Azure Stack in countries where it operates but public cloud Azure is not available, for example in Russia. Carnival Cruise Line is using Azure Stack for private cloud deployments on its cruise ships that operate in periods without connectivity while at sea, particularly in stormy weather. In each case, the I&O team can develop deployment and management methods for applications that apply uniformly across each instance regardless of whether it is running in Azure public cloud or an Azure Stack-based private cloud. Azure Stack is most directly appealing to I&O leaders in companies already using Azure public cloud that want an option for deployment back in private cloud, but it may also drive further preference towards Azure public cloud over Amazon Web Services (AWS), Google Cloud Platform, etc. for its natively integrated hybrid orientation. Its obvious constraint is that it delivers hybrid cloud only in the scope of Microsoft Azure without intent to serve as a cloud management platform extending to management of other cloud providers and platforms like AWS, Google, OpenStack and Dell EMC’s VMware.
One of the leading enterprise vendor candidates to provide hybrid capabilities across multiple different public and private cloud platforms is Red Hat. At Red Hat Summit 2017, Red Hat announced it will soon deliver integrated access to AWS services in its OpenShift Container Platform, which will provide hybrid application deployment capability from a single OpenShift deployment. This will enable I&O leaders with a control system they can deploy in a public cloud or a private cloud as a cloud management platform that controls application deployment by their business units across both public cloud (in this case AWS) and private cloud (e.g., OpenStack or VMware) using Kubernetes-orchestrated Docker containers. Red Hat should be able to build out its platform with strong unified security and spend policy enforcement on top of application deployment and orchestration across both public and private cloud infrastructures.
None of these emerging options for 2017 approaches a full solution to enterprise needs like those enumerated by I&O leaders in ONUG for hybrid cloud as yet. But each offers much bigger steps toward true seamless hybrid cloud than achieved in years previously. Each has tradeoffs based on its technical approach. Microsoft’s focus on its Azure platform won’t address I&O leader needs across other cloud platforms at a time when most enterprises have already adopted at least two public clouds across different application and business unit needs. Red Hat’s control of multiple infrastructures is performed through its interface, so it won’t be as native an experience for AWS-experienced public cloud users as Microsoft’s will be with the Azure and Azure Stack combination. VMware previously announced intent to deliver a similar hybrid capability with its VMware Cloud on AWS. Hewlett Packard Enterprise (HPE) and IBM are certainly interested in delivering value in this space, with HPE due to provide an update on its capabilities and future plans at its Discover 2017 conference.