CIOs are asking their IT teams to build private clouds that reflect the simplicity and agility Amazon Web Services delivers in its public cloud services, including such core capabilities as self-service catalogs, on-demand provisioning and chargeback.
But there's one big problem: Many enterprises struggle to build true private clouds because most IT organizations aren't structured in a way that facilitates such a radically different delivery model.
Cloud providers have an opportunity to help customers develop a new organizational approach that puts a private cloud within reach. But why should providers care about helping customers build private clouds when there's money to be made in public cloud services? The ultimate goal for many enterprises is to interconnect public and private clouds via cloud connectors to create hybrid clouds that are more scalable, elastic and cost-efficient than private clouds alone. In fact, according to Gartner, 70% of enterprises will be pursuing a hybrid cloud strategy by 2015. Customers will need solid footing in the private cloud to achieve that.
However, 451 Research reported earlier this year that 61% of companies were still at the virtualization-only stage in the path to cloud readiness, falling short of a fully automated and orchestrated cloud infrastructure -- two key enablers of the private cloud.
In particular, 74% of the respondents surveyed said they faced "people and process" barriers -- internal politics, budget, time and staff resources -- in making the transition from virtualization to the software-defined data center, self-service clouds and IT as a Service (ITaaS). Cloud providers can position themselves as the trusted advisors that can help customers overcome these barriers and develop a more progressive IT organizational structure.
So, why the delay?
Most enterprises start their path to private cloud adoption by extending their virtual infrastructures with tools that automate orchestration and operating-system provisioning. These are great places to start, but they aren't sufficient to deliver on the promise of a "five-minute virtual machine (VM)," a term coined by John Treadway of the Boston-based consultancy Cloud Technology Partners.
Most IT organizations are still stuck with manual approval and change-management processes that add days or even weeks to the time it takes to spin up a fully functional VM that can be used in production with all of the components it requires. Those components include networking and storage, plus critical governance and IT operations management functions, such as security and compliance monitoring, data backup, server monitoring, configuration management and application performance monitoring.
These functions are often handled by multiple specialized sub-teams of the IT organization in charge of individual domains -- the networking team, the storage team, the monitoring team, the automation team, etc. These teams don't necessarily coordinate their efforts, and they often tightly control access to specialized resources, such as domain-specific management consoles from vendors such as BMC, CA, Cisco Systems, EMC, HP, IBM Tivoli, Microsoft and Splunk.
This organizational model might have worked well in conventional data centers -- where servers changed relatively infrequently and there were far fewer servers to manage -- but it simply doesn't work with the massive scale and dynamic nature of the cloud.
The self-service cloud and ITaaS requires customers to adopt a more vertically oriented IT organizational structure in which a single team can easily administer and monitor the entire stack.
Three IT organization tips cloud providers should promote
When enterprises build private clouds, they will look to cloud providers -- their domain experts -- for guidance not only on technology choices, but also on how to overcome any "people and process" issues to fully realize the benefits of the cloud.
As a cloud provider, here are three tips you can share with customers to help them adopt cloud faster and smarter:
- Apply a DevOps approach to IT operations management. The DevOps philosophy, which supports tightening the collaboration between development and IT operation teams, has been widely adopted by the most successful Software as a Service providers that often make code changes several times per day. This accelerated approach to continuous deployment mandates new methodologies to streamline the entire development process -- from feature definition to code development and testing to going live on the site. Many of these ideas will eventually carry over to traditional IT organizations that aspire to deliver the same level of agility as their public cloud provider counterparts.
- Embrace software-defined everything. In a software-defined data center, virtualization and abstraction extend beyond servers to other parts of the IT environment, including networking, storage and IT operations management. When combined with the development of new application programming interfaces to control each of these layers, system administrators can automate infrastructure changes via programming for the first time -- rather than via manual processes or standalone scripts. This makes the five-minute VM a reality, rather than an unfulfilled promise.
- Think about agility, even after initial provisioning. It isn't sufficient to simply provision a VM quickly -- even if it's a fully featured VM that's ready for production use immediately. The same agility is required on day one for managing ongoing production operations, including such critical lifecycle functions as health monitoring, remediation, patching, configuration auditing and resource management (CPU, memory, I/O, etc.). Fortunately, virtualization and encapsulation make it much easier to instrument, analyze and control each layer of infrastructure so they can be continually tuned for more efficient operation.
The way we deliver information and services via the cloud is undergoing radical change, which requires corresponding changes in IT organizations, processes, technology and the IT organizational structure. Enjoy the ride as you help enterprises navigate these exciting changes.
About the author:
Phil Neray is vice president of cloud strategy at Intigua, a Newton, Mass.-based company whose flagship product virtualizes the IT operations management layer -- just as servers, storage and networks have been virtualized -- bringing simplicity and agility to IT operations for private and public clouds. Its platform, which won Best of VMworld 2012 in the "New Technology" category, is delivered as a subscription-based, on-premises virtual appliance managed via a cloud-based console.
Neray previously worked for IBM, Guardium, Symantec and Alias Research, and started his career as a field operations engineer with Schlumberger Ltd., where he worked on remote oil rigs in South America.
He has a bachelor's degree in electrical engineering from McGill University, where he graduated with honors. He has also earned a Certificate of Cloud Security Knowledge from the Cloud Security Alliance (CSA), and he holds a 1st kyū brown belt in American Jiu-Jitsu.