For most IT organizations, deploying a successful enterprise hybrid cloud is the next step to bringing together all the efficiencies and capabilities they’ve achieved through infrastructure virtualization, standardization and consolidation, and the ongoing evolution of software automation to deliver self-service capabilities.
At EMC IT, we are in the midst of this hybrid cloud transformation, beginning with an internal hybrid cloud platform, called Atlas, which has been providing agile, on-demand infrastructure (IaaS) to our IT users over the past year.
While our enterprise hybrid cloud is continuing to evolve and grow, I wanted to share some insights with you on our project goals, as well as technology and business choices for this important leg of our IT transformation journey. (For more details check out our white paper and reference architecture, EMC IT Enterprise Hybrid Cloud.
Defining your cloud
When our IT executives came to our architecture team meeting a year and a half ago and asked us to build them an enterprise hybrid cloud, nobody really knew exactly what that meant. Hybrid cloud can mean different things to different users. We knew we wanted a more agile, flexible way to deliver IT services without the constraints of physical and on-premises data centers. But we really needed more specifics to come to a common understanding of cloud computing for our transformation project.
We began by researching definitions of cloud and found that the National Institute of Standards in Technology (NIST) definitions were neatly aligned with what we were trying to do.
NIST defines cloud computing as “a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”
NIST also has five characteristics of cloud computing that we decided to use to guide our design: on-demand self-service, resource pooling, rapid elasticity, measured service and broad network access.
Once we had our design principles, we looked at our existing data center to see where we could be more efficient, agile and less error-prone, and more relevant to the business. We found two key issues—our ongoing propensity for custom-designed environments was inefficient and required substantial resources to maintain, and the fact that we were still somewhat siloed as an organization was inhibiting quick reaction time and preventing the pooling of staff resources.
We set out to get rid of customization and drive up server utilization. We reviewed our most common requests for services, found common denominators for what we were doing and came up with ways to deliver those common requests as standard services through software defined infrastructure. No more building “snowflake” solutions that are difficult to expand, manage and scale. We moved to a standard environment, with standard infrastructure and repeatable processes.
After building our internal cloud, we needed to extend our computing capabilities to include off-premises cloud infrastructure to provide the flexible data center capacity to allow us to rapidly scale.
Choosing a cloud solution
There are evolving enterprise hybrid cloud solutions that provide the technology stack to help organizations create a software defined data center and extend IT capabilities to include off-premises cloud resources.
We found EMC’s Enterprise Hybrid Cloud solution and converged architecture met our hybrid cloud needs. EHC’s fully integrated technology stack provided virtual networking, storage and security, in addition to the virtual compute layer.
The integration of this stack with a public cloud offering from VMware (vCloud Air) further enhanced agility and cost effectiveness by providing on-demand capacity to avoid the inherent delays of adding capacity to our internal cloud. (See more details on EMC IT’s Enterprise Hybrid Cloud Strategy.)
We went live with project Atlas in October 2015, providing more than a petabyte of VMAX storage to users leveraging automation. We have also created more than 250 virtual machines to date and delivered F5 automated load balancing administration to our main internal IT customers—the managed hosting group, the middleware team and the architecture team.
The result has been efficient, fast and easy access to infrastructure at the touch of a button for these users. Previously all infrastructure requests funneled through managed services, which processed the request and created the machines. Now each of these groups has their own portal where they can hit an “easy” button and create their own infrastructure without requiring other groups to get involved. And they can do it within minutes rather than days or weeks as they did in the past.
Highlighting the possibilities
Since Atlas is an iterative project, we are still working to deliver additional services and components for the managed hosting team. And now that we have built the foundational cloud platform, we are able to leverage it to add more services to the business as they come to us with requests.
With the hybrid cloud, it doesn’t matter to users whether their needs are met by on-premises or off-premises resources. The bottom line is that IT is able to show business value by using hybrid cloud technology to enable users to access environments quickly and securely with consistent performance, lower error rates and lower costs.
Of course, transitioning to an enterprise hybrid cloud platform requires a major cultural adjustment to get people to make the process change needed. We are continuing to evangelize our solution, and act as influencers to the organization to show users the tremendous potential of our new cloud factory.
For more information be sure to check out EMC IT Proven’s Enterprise Hybrid Cloud Reference Architecture.