“Having a datacenter seamlessly extended to a single public cloud offering is an interesting idea. But a “One cloud” approach isn’t going to fit the reality of where the industry is headed.” Gaurav Maglik, CEO CliQr Technologies
Building on a key idea from CliQr CEO Gaurav Maglik’s recent Cloud Strategy interview.
“One Cloud” Incumbent Investment
I understand why many industry incumbents have a hybrid cloud offering that fits their business interests. VMware’s “One cloud, any app” solution connects vSphere customers to vCloud Air, the public cloud service they have heavily invested in. Microsoft Azure Pack can be used to connect Microsoft Systems Center to Azure, where Microsoft also has a large capital investment. Red Hat has refocused Cloud Forms to connect OpenStack to one cloud, Amazon Web Services.
But we are only in the 2nd or 3rd inning of the cloud game. IT executives are nervous about picking winners this early in the game, and are actively working to avoid cloud lock-in. If we polled 500 CIOs and asked, “In 5 years – do you think you will be managing your datacenters and one public cloud, or an evolving and agile mix of your datacenters and multiple public clouds, where IT helps find the best fit for business needs?” I’m guessing the most common answer is not “one cloud”.
Multiple Cloud Reality
In fact, we already see global 2000 companies dealing with the reality of managing and integrating processes across multiple clouds. We have a customer supporting one development group that is standing up OpenStack to spin up dev and test environments on demand. Another dev group in the same company is using AWS for the same type of work. But IT Ops has a large VMware vSphere footprint. They need to deploy and move applications across these environments to support a diverse mix of business needs.
How does an IT executive or enterprise architect make sure these diverse cloud efforts are secure and compliant? How can you avoid spinning up multiple new silos of cloud teams and management stacks, at a time when IT is facing a skills shortage? How can IT efficiently and securely support diverse cloud efforts that each have solid business justifications?
Common Design Patterns
Trace3, one of CliQr’s top partners, conducts Cloud Expedition Workshops with a wide variety of Global 2000 clients. Anoj Willy, Trace3 Vice President of Cloud Solutions, has found three common design patterns as a result of analysis of business needs and specific cloud solution scoping with multiple customers:
- Instant Developer Environments. Give developers the ability to spin up development and test environments without an IT help desk request. They want more than infrastructure on demand. They want fully configured application stacks, including middleware, as well as development and test tools. Developers like instant access in public cloud, but have found that they spend a good deal of time creating their work environment on top of the instant infrastructure. Deploying the entire application stack so a developer can begin coding or testing immediately, without installing and configuring software, is the goal.
- Multi-Cloud Application Profiles. Create a profile of your application and it’s major components that you can deploy, version, and maintain across multiple cloud platforms both private and public. Giving on demand, self-service access to application stacks across multiple private and public cloud environments allows users to work directly on the application content down through its infrastructure without needing to know the nuances of each cloud target. Users want a simple user interface and SLA guarantees around their code deployments. The policies that dictate security, performance, and cost should be packaged into a simple offering so that users understand the tradeoffs and can select the optimal choice from available environments.
- Automated Integrated Lifecycle. Once Dev and Test stacks can be deployed on demand, the next step is integrating tools like Jenkins, Git and Docker to automate code deployments, build and test. This isn’t lights-out, full-process automation. This is an integrated tool chain that eliminates process white space, speeds up deployment, and squeezes out additional non-development work. This may be automated across environments where dev and test, staging and production can be deployed on different datacenter and public cloud environments.
The point is, management stacks optimized for incumbent vendor business interests may add value in one cloud environment, but they aren’t designed to support these real-world design patterns that include multiple clouds to meet varied business needs.
A “one cloud” strategy in a multi-cloud reality creates new silos with cloud-specific teams, skill sets, as well as automation and orchestration solutions that need to be repeated for each cloud.
CliQr Federated Cloud Approach
A federated cloud strategy offers a better approach. CliQr CloudCenter is a single platform that is designed to easily deploy and manage application stacks on any supported cloud. We just added support for IBM Softlayer and can now manage applications on over 20 datacenter and public cloud environments.
- The management features allow fast modeling of application stacks and automated deployment to give users what they need, when they need it.
- The orchestration capabilities eliminate the need to code applications or create deployment scripts that are specific to a single cloud environment.
- Governance features give enterprise IT organizations single pane of glass visibility and control across multiple clouds. Tag-based governance and security profiles guide automation to optimizes workload placement and actually increases security and compliance control over less automated approaches.
Don’t bet on a “One Cloud” future. Don’t get locked into a single management stack.
Instead, “future proof” your cloud strategy and use a cloud and application management platform that supports a federated cloud approach and meets a wide range of needs today and in the future.