Bringing Order to Chaos through Data Center Automation

InfoManagement Direct, October 2007

Jennifer Ellard

When Henry Ford installed the first conveyor belt-based assembly line in 1913, the car industry was forever changed. Enabled by this streamlined process, Ford workers could assemble the famous Model T in just 93 minutes - a dramatic improvement over previous production methods, which took almost 3 days to assemble one car. In addition, Ford only offered the Model T in one color - black, invoking a theme of consistency and standardization very early on in the industrial era.

Nearly 100 years later, enterprise data centers are on the verge of experiencing their own version of the automotive revolution. IT managers are tasked with effectively managing the infrastructure while addressing several operational challenges that contribute to inefficiencies in the data center. Many of these challenges can be traced to the following three factors confronting IT today:

• Stringent availability requirements overlaid with “always-on” expectations. The bulk of the business world is carrying around a BlackBerry, Treo or some other handheld device that gives them the ability to respond to requests instantaneously. This created an expectation that everything must be available all the time, and that expectation has been passed on to IT infrastructure availability requirements. It’s the pace of the world that we live in today. More importantly, the value of information to the business has increased dramatically, as well as the sheer number of transactions processed. Consequently, the cost of downtime has skyrocketed and service level agreements (SLAs) have risen in stride to keep pace. This has further pressured IT to minimize downtime and drive efficiency wherever possible.
• An increase in data center complexity. Over the years, the IT environment has become more complex due to the growing number of multitier applications, heterogeneous platform environments and the introduction of virtualization. Simply put, there are more moving parts in the data center, and it’s become increasingly difficult for IT managers to get their arms around the complexity and manage it effectively. They lack the visibility of what is running in the environment and how it’s all tied together. This lack of visibility has also increased change-related downtime and overall business risk and has decreased IT’s ability to implement changes rapidly and respond to the business.
• Tight resource constraints that hold IT budgets and headcount flat. Large-scale enterprises have demanded that IT evolve from a traditional cost center to a source of strategic and competitive advantage that is cost-efficient and dynamic. This “do more with less” mandate has forced IT to rethink its approach to management and look to methods for optimizing its operations across the infrastructure.

Within these facts lies the challenge of managing today’s enterprise data centers, which is driving IT organizations to look to the assembly lines of traditional manufacturing to apply three key principles: first, keep the assembly line running at all costs; second, identify and automate high-volume repetitive tasks; and third, leverage the resulting innovation to increase productivity. In short, drive higher availability, automate IT processes, and leverage the staff and infrastructure utilization. Data center automation enables IT management to automate labor-intensive data center tasks, with the goal to enhance visibility and control, reduce errors, increase service levels and drive down cost. IT departments that do not take advantage of innovative data center automation (DCA) solutions will continue to struggle with operational inefficiency and find themselves on a collision course with chaos.

Strategic Challenges in Addressing Data Center Complexity

As data centers have coped with intensifying pressure to manage more infrastructure with less resources, a number of potential, nonautomated strategies have arisen delivering some significant benefits, but they are not without costs.

Many enterprises have chosen to go down the path of server, application and data center consolidation to achieve operational efficiencies and streamline management tasks. In fact, much of the rise of virtualization can be attributed to the desire to drive up server utilization and enhance the flexibility of the infrastructure.

While consolidation and virtualization both have their merits, these technologies also introduce a fair amount of risk to the IT infrastructure given today’s complexities. For example, many enterprises have experienced outages during their application migration or consolidation efforts due to unknown dependencies across the infrastructure. Virtualization drives up utilization and improves dynamic mobility, but also increases the number of users per server, potentially creating broad-scale single points of failure.

Add heterogeneous platform OSes and a lack of visibility into the mix, and the data center becomes a murky witches’ brew of management tools with limited scope to control the enterprise. The challenges become more apparent as enterprises throw more server horsepower at the problem, creating scale-out server models or virtual machine sprawl to accommodate business demands.

To solve these issues, some organizations have chosen to leverage customized management solutions to rein in data center complexity. However, these toolsets require dedicated development cycles and often do not account for next-generation data center architecture, leaving them subject to obsolescence or redesign, or limiting IT’s choices of future technologies. Other IT shops select vendor-specific framework tools, which lock the enterprises into inflexible processes based on the provider’s technology and create additional costs downstream.

Data Center Automation: A Strategy is Required

As technology options improve, the most forward-thinking companies are carefully assessing these options and are increasingly taking a strategic approach to data center automation (DCA). This approach entails looking at the challenges that need to be addressed today from an operational efficiency standpoint as well as planning for tomorrow’s architecture and the management toolset that will be required.

In assessing suites of DCA products, some of the most critical capabilities to look for include centralized management capabilities and support for heterogeneous systems. In addition, the ability to manage across traditional IT silos and provide workflow capability across data center infrastructure will become more important as the environment becomes more dynamic. With the rise of virtual technologies, both server and application-based, it will be critical to have management capabilities to control the physical, virtual and application infrastructure

Furthermore, enterprise scalable availability technology will become even more critical as higher utilization will essentially put more eggs in one basket for IT and increase the level of risk. Organizations today are already spreading virtual machines out, reducing utilization to disperse the risk element. A pooled architecture and technology to dynamically support it could enable consolidation and efficiency of a previously unrealized scale.

Steps Toward DCA: A Logical, Measured Approach

As such, any strategy toward DCA should move beyond examining the management capabilities gained today to also include how automation and computing architectures will shift going forward. A sound DCA approach will look not only at automation, but also at how to address the gaps and problems potentially created tomorrow - how to enable dynamic resource allocation and drive responsiveness and flexibility to the business needs.

First and foremost, IT organizations will want to identify external frameworks that can alter data center management requirements, such as Information Technology Infrastructure Library and service-oriented architecture. Any technology solution should work to augment these frameworks as well as any processes that are instituted as a result.

After the IT organization has agreed on the standards of success, it should determine the precise measurements that constitute optimal resource use. For example, measurements can include the DCA products’ compliance with ITIL and ISO standards, internal best practices (existing or to be created) and standard operating procedures, and the ability to meet internal service levels.

Taking a comprehensive approach to achieving DCA uniquely holds the most power and potential to transform data center operations. IT organizations can follow three important guidelines to help ensure a holistic approach:

1. Develop a strategy with objectives before investing in any data center automation technology. All data center systems should be included within the resulting plan.
2. Avoid point products that do not provide centralized management and automation capabilities. IT organizations should invest only in solutions that assist in automation of all data center systems and tasks.
3. In choosing which areas to address, IT organizations should seek to automate repetitive, manual tasks and increase the utilization of hardware and staff resources, focusing first on the most valuable projects.  

Leveraging DCA to Transform IT

Next, the IT organization should seek to leverage its DCA solution to drive operational efficiency and responsiveness to the business based on a tiered maturity model. The maturity model determines the appropriate fit and implementation level of a DCA initiative. A tiered model in Figure 1 represents the building blocks for an organization to target.

• Availability,
• IT process automation, and
• Policy-based control.

Within this model, at a fundamental level, availability represents the data center’s basic needs. To use an analogy of Maslow’s hierarchy of needs, availability is the food and water for the data center - the most basic needs, because if the systems aren’t available, IT has failed to deliver to the business.

Process automation represents the next tier - we’re starting to automate some mundane and repetitive tasks to achieve efficiency and cost savings and to leverage staff further. In the analogy, the data center has moved beyond basic needs to put a roof over its head; maybe it has a couch and a feed for cable TV.

Policy-based dynamic control represents the tier that will provide IT with the greatest returns for the company. It’s the central AC and heated bathroom floor. It’s what gives data center managers the ability to control changes and move applications throughout the environment, consolidate and expand based on utilization. This tier enables policy or runtime-oriented responsiveness and the automated capability to deliver against SLAs. It gives you visibility and control to optimize the environment in spite of the complexity that exists.

Bearing these guidelines in mind, the next step is to select a DCA solution that is fundamentally holistic. Specifically, the solution should be envisioned by the provider and architected to be an end-to-end offering, encompassing servers, applications and storage. Enterprises should expect a solution for DCA that is both complete and simple, as opposed to accretive and complex.

Similarly, the DCA solution should not be limited to a single tier of the architecture. It should automate the management of the server and application tiers of the data center, providing integrated, multilevel application availability. The DCA solution should also deliver automated configuration and provisioning, which in turn can give IT organizations better visibility into change, with improved impact analysis.

Automation Benefits for the Next-Generation Data Center

By adhering to these guidelines, choosing and properly deploying a next-generation DCA solution, IT can better control data center infrastructure, enabling wide-ranging improvements that drive business benefits. These include:

1. Business service availability: Reduces planned and unplanned downtime with automated failover and recovery, impact analysis and configuration rollback.
2. Improved change management: Provides impact analysis, automates deployment of changes to servers and applications, and maintains change audit trails.
3. Increased operator productivity: Enhances operator efficiency by centralizing application management across heterogeneous environments and automating routine provisioning, patch and configuration tasks.
4. IT policy and standards compliance: Defines deployment standards; supports image-based deployment and server comparisons and monitors and tracks drift from those standards over time.
5. Comprehensive high availability and disaster recovery: Protects against local, metro and regional outages with automated recovery of complex, multitier applications. Dynamic environment sharing provides cost-effective protection for less critical applications.

In sum, as today’s data centers struggle with stringent service level and availability requirements, infrastructure complexity and tight resource constraints, companies should consider data center automation strategies that improve the way the IT infrastructure is managed and administered. By implementing a next-generation DCA solution, IT can realize the transformative power of automation as a framework to enable dramatic improvements in the efficiency and effectiveness of data center operations - not unlike the introduction of the assembly line at Ford Motor Company nearly a century ago.

Jennifer Ellard is a senior product marketing manager with Symantec's Data Center Foundation.

For more information on related topics, visit the following channels:

• Best Practices/Benchmarking
• Data Management
• Data Management