Andrew Hillier

Internal clouds are gaining momentum in large organizations, both as a means to drive agility and efficiency and as a stepping stone to hybrid and external cloud models. Although establishing and managing an internal cloud seems simple and elegant in principle, in practice they rarely resemble the simple, fluffy white cartoons seen in presentations. More often than not, today's cloud strategy involves piecing together hypervisor management layers with self-service portals, orchestration solutions, provisioning tools, and other base components. While this will yield a functioning cloud, it tends to operate at a relatively basic level, with little intelligence powering its action. This is a lot like having a body with no brain: if you tap its knee it might kick, but it is only capable of basic, reactive behavior.

Some might argue that this is no different from traditional IT physical environments, where disjoint tools and processes can be effectively used to handle specific aspects of operations and management. But the difference in these new environments is that there is a much tighter inter-relationship between all of the physical and virtual components. Physical complexity is replaced by virtual complexity, and characteristics like multi-tenancy and resource overcommit add a new twist. All of these factors combine to create an interesting challenge in the management of these environments, taking it beyond the realm of what intuition, spreadsheets and simple load-balancers can effectively handle.

In order to host cloud workloads in an efficient and low-risk way, the equivalent of a brain is needed. Like a human brain, it must aggregate and correlate information regarding the configuration and operation of the components in the system in order to maintain a certain level of "situational awareness." But this alone is not sufficient, as it provides visibility but not action, and is reactive, not proactive. To safely, reliably and proactively manage cloud environments there are two equally critical inputs that are required: cloud management policies and future capacity bookings.

Cloud management policies are the rules, constraints and preferences that govern the relationship between capacity and the workloads that run on it. These policies are critical to cloud management, and affect everything from utilization levels to resource overcommit, technical compatibilities and business rules. Interestingly, most of these considerations have existed for quite some time in IT environments, but there was less of a need to formalize them into a true policy until now. This is mainly because physical environments are not flexible enough to warrant it, and early virtual environments have targeted the "low hanging fruit," typically at the departmental level, which tends to be comparatively simple to manage. But a cloud with no policy is not viable, as it by definition hosts disparate applications from different business groups, each having different performance, availability and security requirements. Throwing these together randomly is a recipe for failure, and can lead to many performance, compliance and efficiency issues.

Once policies have been established, the other key ingredient is a representation of the future capacity bookings. This is the piece of the puzzle that allows an organization to attain a level of predictive analysis, which is essential to proactive management. These bookings can be thought of as hotel reservations, and represent the up-coming demands related to new application deployment, migration and consolidation activity, and other requirements. While most think of clouds as a way to get capacity "right now," in many organizations the goal is more strategic than tactical, and booking capacity with some level of advanced notice is far more consistent with the operational model. This provides tremendous advantages in terms of agility by removing the procurement cycle from the critical path, while also providing a level of control over IT assets. Immediate requests can still be accommodated, of course, but may incur a premium due to the volatility they introduce.

With these inputs in mind, let's get back to the brain. Data gathered from the "eyes and ears" of the environment (e.g., monitoring and alerting tools) provides situational awareness through a knowledge of current and historical activity, configurations and status. Leveraging a brain, or applying analytics to interpret this information in the context of the prevailing policy for a cloud or specific workload within it, enables risks to be identified and efficiency quantified, thus providing judgement as to whether action is needed to improve the environment. Factoring in future bookings enables "future situational awareness," where policies can be used to drive predictive analysis of potential risks and inefficiencies. Just like the brain, these current and future actions can then be communicated to the "arms and legs" of the environment (e.g., management, provisioning and orchestration tools) to achieve the desired outcome.

The benefits of a more intelligent approach to cloud management can be staggering. Isolated and reactive tools, however well integrated, are simply not capable of managing IT to a high degree of efficiency. But the combination of these tools with proactive, policy-based analytics provides significantly greater control, yielding higher efficiency while at the same time significantly reducing operational risk. In today's virtual and cloud environments, this translates into better performance and significantly higher density, saving millions of dollars in infrastructure costs.