Intel DCM and iTRACS: Towards a New Efficiency Metric

Introduction
Data center infrastructure management (DCIM) technology has made great strides in recent years, providing ever-more granular views into the use of electric power and computing capacity in data centers. With the rising cost of energy and the complexity of managing the explosion of new data sources and mobile applications, data center operators are under pressure to improve efficiency, agility, and dependability. Each iteration of DCIM provides another angle from which to understand the relationship between energy consumption and computing power. Moreover, business leaders want to understand an even more elusive relationship—the relationship between power consumption and business value.

A particularly ambitious DCIM provider, iTRACS already has a 360-degree view of DCIM—literally. iTRACS offers interactive 3D visualization, a navigable point-and-click 3D environment that lets users see, manage, and gain complete “command and control” over physical infrastructure in the data center, from the transformer outside of the building down to every device on every rack.

Here are two real-world examples of iTRACS 3D visualization software in action:

Now, by integrating Intel Data Center Manager (DCM) software with its Converged Physical Infrastructure Management™ (CPIM™) software portfolio, iTRACS is able to extend its DCIM capabilities even further, along two lines:

• iTRACS can now measure the power consumption and environmental conditions of each asset in the data center, right down to the device level.
• iTRACS is working to develop a new “business value” efficiency metric that would augment PUE by calculating the efficiency with which IT assets—the compute side of the data center—deliver business value to the organization. Essentially, the more business value delivered by IT per unit of energy consumed, the more efficient the data center becomes.

About iTRACS

iTRACS, based in Oak Brook, Illinois, has been developing physical infrastructure management solutions for 20 years, beginning with building cabling infrastructure in data centers. The company moved from asset-management documentation in the late 1980s, to embedded resource-discovery tools in the early 2000s, to automated Move/Add/Change platforms in the middle of the decade. The CPIM™ software portfolio with interactive 3D visualization was launched in 2009, providing users with a real-time, navigable 3D environment for managing all data center assets and the complex web of interrelationships that exist between them.

iTRACS has built an energy efficiency strategy around its CPIM™ software portfolio called iTRACS PowerEye™, which consists of four main energy management capabilities: Visualize, Analyze, Manage, and Optimize. By “flying” around a 3D model of a data center, an operator sees a real-time picture of the entire environment, clicking on assets to reveal which applications are running where, what each network cable is connecting, and which assets are underutilized or wasting power. This level of point-and-click navigability offers true command and control over the physical infrastructure.

“If the boardroom tells you that you need to cut energy consumption by 20% by consolidating servers—while at the same time increasing the number of virtual machines on the floor to optimize your virtualized environment—how would you figure that out?” asks Willie Bloomstein, Marketing Strategist, iTRACS. “Previously, you would send out a team with clipboards to walk the floor, trying to figure out which servers were connected to what and where space is available to replace older devices with new energy-efficient ones. Now you can sit at a console showing you a single-pane view of the entire data center, examine your environment, explore ‘what if’ scenarios, and come up with a strategy that mitigates the risk associated with change. It’s not just real-time event management or reducing energy consumption. It’s also the ability to strategically optimize physical infrastructure over the long term to deliver continuous business value to the organization.”

iTRACS estimates that its iTRACS PowerEye™ strategy can help companies, depending on their current power usage, reduce their energy costs by up to 30 to 40 percent.

Role and Benefits of Intel DCM

Intel Data Center Manager (DCM), completes a vital link in the iTRACS campaign to provide a total picture of the business value generated by IT-related energy consumption. With the integration of Intel Data Center Manager (DCM), into the CPIM™ platform, iTRACS extends its capabilities deeper into the data center, with the ability to collect, manage, and analyze power and temperature readings at the individual device level. Leveraging this granular information, data center managers can improve capacity planning, identify and decommission energy-wasting assets, strategize new equipment outlays using predictions based on actual energy usage, and prevent outages from occurring by identifying and resolving sources of overload, hot spots, and cooling issues before they become problems.

On the Road to a New Efficiency Metric

In addition to improving energy efficiency, the iTRACS and Intel partnership is bringing the DCIM world one step closer to a new paradigm in the measurement of efficiency. This new approach, which iTRACS calls DCIM Business Output™, is designed to augment the popular PUE metric with a series of new measurements that calculate the efficiency with which IT assets—the “compute” side of the data center—deliver business value to the organization.

Explains Bloomstein, “PUE is an excellent guide for measuring the percentage of energy going to facilities assets (like cooling) vs. the amount of energy going to IT assets (such as servers). But PUE is facilities-centric, telling you little about the efficiency with which IT assets are driving business value for the organization. We are suggesting that the industry supplement PUE with an additional IT-centric approach, which would measure the positive business outcomes delivered by the IT assets, and how much power is required to achieve those outcomes.”

Asset Efficiency

This proposed metric would compare CPU utilization vs. power usage at the individual device level, creating an opportunity to compare one IT asset against another to see which provides the most value at the highest efficiency.

Comparing Server A with Server B might look like this:

Server A

• Capacity: 100 business operations per minute
• Is currently running at 87 operations per minute—operating at 87% capacity
• Consumes 100 units of electricity—running 0.87 operations per unit of energy

Server B

• Bigger capacity: 150 business operations per minute
• Currently running at only 75 operations per minute—operating at 50% capacity
• Also consumes 100 units of electricity, so it’s running 0.75 operations per unit of energy

So server B, which has the potential to deliver higher business output to the organization, is actually running less efficiently. Both the business and the planet suffer.

IT Efficiency

Bloomstein calls this second metric the “PUE of IT” because it would help determine how much energy is going into IT assets that work directly for the business (servers, etc.) versus IT assets that do not work for the business, but are in a supportive role (storage, switches, etc.). The goal is to minimize the amount of energy consumed by assets that aren’t working directly for the business so that owners can put more energy into assets that are doing work for the business.

Bloomstein says the “business value” or “business output” in these metrics would be whatever each individual data center operator wants them to be, depending on the organization’s business model and mission. It could be as simple as CPU cycles or as targeted as customer transactions. An ecommerce company might measure “business output” as transactions or revenue. A social media company may look at advertising dollars or social transactions. A cloud provider might look at services or compute power charged to customers.

iTRACS believes a new family of IT-centric metrics could be helpful for businesses that increasingly run in the cloud, allowing providers to apportion computing resources, in real time, based not only on customer demand, but on climatic conditions and energy costs at data centers at certain points in the day. A services provider with a global portfolio of data centers could essentially move the cloud to different locations on the planet at any given point in time, depending on where the greatest energy efficiency is available. The business would expend the least possible energy while reaching its most ambitious goals.

Conclusion

Like the famous Charles and Ray Eames film The Power of Ten, the vision of iTRACS evokes a similarly giddy level of experiential engagement with the DCIM landscape, moving from a grain of sand (a device) to a view of the whole planet (global businesses running in the cloud) in an instant. The level of detail provided by Intel Data Center Manager (DCM), means that one will eventually be able to jump from the grain of sand into the cloud and clearly see power usage and other key metrics at any scale.

All of this has a lofty goal—a more energy-efficient, optimized business environment that will have a truly granular yet global understanding of the impact of its decisions on its bottom line and on the planet.

Related Articles

Eliminating Massive Amounts of Energy Micro-Waste
High Performance Computing Vendors Gain Energy Advantage
Power Management at the Data Center: From Capacity to Control