Vendor View

Meeting the Need for Unified Virtualization Management

Virtualization and the cloud are all about giving IT the leverage it needs to management today's IT processes.

• By Lauren Whitehouse
• 12/04/2012

Cloud computing has had a huge impact on the entire IT industry by changing the way IT professionals and technology companies develop, implement, and manage IT environments. This impact is significant in the case of workload management, where the sheer scale and complexity of cloud computing environments are causing organizations of all sizes to rethink IT management processes.

The Traditional Approach
Traditionally, IT management has taken a bottom-up approach. We collect information and track metrics on hundreds of things in the data center, and then we use this information to improve infrastructure and operational efficiency. When a metric hits a pre-defined threshold, we're alerted to the out-of-policy condition and an operator must interpret the data and determine the appropriate course of action. In this approach, there is no proactive resolution. It is all reactive and time-consuming because an administrator is required to manually address each problem that is identified.

Decades of delivering volumes of information that require operator intervention and acumen to diagnose and remedy has resulted in the invention of dozens of niche solutions that are used to collect data, provide reports, and alert IT to take whatever action they think is required. The only innovation has come in the form of consolidating and suppressing alerts (filtering to distinguish the critical from the mundane), and "learning" to predict what could happen. To date, these solutions have certainly helped IT meet basic service levels, but with the evolution of virtualization and cloud computing, they could never scale to fit these environments.

Virtualization and cloud computing have made this already clumsy model exponentially more complex and unrealistic to successfully manage. This is because, in virtualization, the causality relation is constantly changing due to the dynamic nature of the environment. Understanding the interdependencies between workloads, systems and resources to pinpoint and resolve issues can be an impossible task with tools not designed for virtual and cloud architectures.

Keeping up with the evolution of computing requires challenging the traditional approach employed by today's virtualization management software. Rather than offering IT organizations more and more point tools that provide little value and many more headaches in the form of meaningless data, endless alerts, more consoles, and manual intervention, there is a real need for unified virtualization management foundations that break down the boundaries and integrate element and hypervisor management, manage all layers of the stack, and proactively adapt resources to assure performance.

New Management Solutions
Trying to address problems after they've occurred is neither a viable nor practical solution. We need to begin taking advantage of the fluidity and control options that virtualization affords. To achieve this we would need an approach that entails a closed loop process that monitors, analyzes and controls the virtualized environment, so that rather than alerting IT managers when problems occur, or are about to occur, these problems are prevented from happening in the first place. The approach would keep the environment running in a healthy state by optimizing workload performance, maximizing resource efficiencies, and reducing operational overhead.

In this new management paradigm, the multitude of niche tools that organizations use to manage their IT environments would become obsolete. In their place would be an integrated management system that would work off the premise that in order to keep the environment running efficiently in a way that proactively solves problems, we have to start by understanding what the infrastructure and operations teams need to know about in the environment, and what they can control.

A Smarter Way
What we need is a new approach to workload management that emphasizes prevention - a proactive system that keeps the virtual data center healthy by continuously tuning the resources under management. The system should ensure that required resources are available to workloads across the shared infrastructure based on workload service level priorities. To achieve this, the systems must automatically and continually assess performance and resource contention. This will give IT the control it needs to meet business objectives and eliminate the reactive, time-consuming, and labor-intensive fire-fighting typical with monitoring-and-alerting approaches.

Managing risk, lowering costs, and increasing agility of IT services are happening by leveraging virtualization and the cloud. However, management deficiencies can result in poor Quality of Service, reactive troubleshooting, and/or diminished utilization rates, detracting from user progress in meeting IT service delivery goals. With a new approach, organizations can maintain the right balance of resource utilization and application performance. This will enable them to deliver enhanced service assurance, improved resource utilization and operational efficiency, less risky SLAs, and lower costs.