BIoT Canada

The Mainframe: A Platform For Energy Efficiency

For comparable levels of computation, mainframe systems take up dramatically less space and are proving to be extremely efficient in cooling and power consumption.

March 1, 2009  

Print this page

My eight passenger car may not initially seem very energy efficient, but I can consolidate and carry many passengers.

I can virtualize by successfully and securely transporting the old, the young, male and female, dogs, telescopes and sports equipment, all in the same vehicle. And suddenly I have a very energy efficient solution.

Energy efficiency for data centres has become a crucial area of focus as the price of energy increases and systems grow beyond the capacity of current facilities to supply their power and cooling needs.

Using energy management strategies including virtualization and consolidation, the mainframe, in some ways like my car, is an outstanding platform for power efficiency. This article will discuss the dynamic infrastructure and energy efficiency of large systems, highlight energy management, and conclude with recommendations for the future.

The Dynamic Infrastructure and Energy Efficiency of Large Systems: The world works in very sophisticated ways these days. Every person, business, organization, government, natural system and man-made system interacts.

Each interaction presents us with an opportunity to do something better, more efficiently and more productively.

Interconnected technologies are changing the way the world works: the way my car, for example, was designed and developed, manufactured, bought and sold; the way services are delivered; the way people, money, water and everything else on this planet is moved. The catalyst for this change is the transformation of the world’s infrastructure, which is becoming increasingly instrumented, interconnected and intelligent.

A dynamic infrastructure will bring more intelligence, automation, integration, and efficiencies to the digital and physical worlds.

It will enable businesses and governments to better manage challenges presented by today’s globally integrated planet.

There are three critical requirements for this new infrastructure: the ability to integrate all of an organization’s digital and physical infrastructures, the capacity to store, manage and analyze massive amounts of data, and the means to reduce inefficiencies in business processes, technology and costs.

This dynamic infrastructure must encompass service management, asset management, virtualization and consolidation, information infrastructure, energy efficiency, security, and business resiliency.

Data centre costs including energy have risen eight times what they were since 1996; average distributed server utilization is just 6-15%. So for a data centre, smarter can involve server consolidation, space savings, and energy efficiency. Smarter can be a lower total cost of ownership. Smarter can be a mainframe.

The mainframe is designed to increase data centre efficiency by significantly improving performance and reducing power, cooling costs, and floor space requirements.

A large system can offer unmatched levels of security and automates the management of IT resources to respond to changing business requirements. A mainframe delivers, in a single footprint, unprecedented performance and capacity growth while drawing upon the rich heritage of previous generations.

A mainframe meets the needs of large enterprises having large scale, mission critical transaction and data processing requirements while also delivering the scalability and granularity to meet the needs of medium sized enterprises.

For comparable levels of computation, mainframe systems take up dramatically less space and are proving to be extremely efficient in cooling and power consumption.

Many enterprise and business class systems even have a hybrid cooling system that is designed to lower power consumption.

Energy management strategies, including server consolidation and virtualization, are also a key component to the efficiency of a large system. The mainframe can be an integral part of this strategy by leveraging strengths such as:

• The ability to host hundreds of workloads on a single server.

• Providing advanced management and automation techniques.

• Offering highly secure and robust technologies.

Virtualization is a key strength of a large system. Virtualization can be supported through both hardware and software. Virtualization creates the appearance of multiple concurrent servers by sharing the existing hardware.

Its major goal is to fully utilize resources, lowering the total amount of resources needed and their cost.

Distributed servers often have low utilization levels, which makes them a prime target for consolidation on a larger server using virtualization and workload management. The mainframe runs more application environments on fewer processors and has sustained high utilization levels without reducing throughput and response times.

Consolidating low utilization servers on a large system or mainframe will help reduce power and facility costs.

Energy Management on the Mainframe: Several tools are available to monitor the power consumption and heat dissipation of a mainframe. An estimation tool, an external power meter, management console, and specialized energy efficiency software can assist in measurement and monitoring of systems.

To assist in energy planning, a power estimation tool can estimate server energy requirements before a new server purchase. A user inputs the machine model, memory, and I/O configuration and the tool will output an estimate of the system total heat load and utility input power.

For planning and performance purposes, an external power meter can be very useful in measuring and monitoring a large system.

Actual power consumption of the system can also be seen on a management console. A “mainframe gas gauge,” which my car also happens to have, can provide power and thermal information. Current total power consumption in watts and BTU/hour as well as the air input temperature are useful metrics.

Advanced software tools can help IT managers control and automate large numbers of physical and virtual servers across a full range of hardware.

Through a single interface, users can map virtual resources to physical servers; throttle energy consumption up or down as needed; and collect data on hardware temperatures and data centre energy use.

This type of software can automatically monitor remote hardware operations and take proper action based on alerts.

Software tools can provide a single view of actual power usage across multiple platforms and increases energy efficiency by controlling power use across the data centre. They can provide visibility to and enable energy savings and cost reductions. This type of tool can enable organizations to:

• Increase energy efficiency by monitoring power use across the data centre

• Negotiate the best utility rates based on accurate trend assessments

• Limit server power requirements by capping maximum power consumption

• Manager power use and potentially reduce power costs

• More effectively plan new data centre construction or modifications

• Plan power capacity requirements based on actual usage

• Justify incremental hardware purchases based on available power capacity

• Better utilize existing resources

The Future of Energy Efficiency on the Mainframe: The mainframe provides sustained leadership in energy conservation and management by continuing to deliver power-management and cooling technologies. With these technologies, systems use less power, generate less heat and use less energy to cool the system.

For the future, we should expect to see even more efficient server design. We should see increased server energy efficiency techniques available such as power capping and specialized power napping of system components. Performance benchmarks will incorporate even more sophisticated energy efficiency metrics.

And we will see an extension to energy efficienc
y tools on the server which will bring automation to the management and reporting of energy consumption to non-IT assets — an office building air conditioning system, for example, or streetlights in a city.

With this new type of software, organizations will be able to generate reports to help them track and visualize energy dynamics.

They can then take appropriate action while extrapolating how changes will yield different business outcomes using sophisticated “what if” calculations.

Through innovative technologies, energy management strategies such as consolidation and virtualization, and energy management tools and techniques, it is apparent that the mainframe is a truly outstanding platform for energy efficiency and will continue to be so in the future. Just maybe, like my car.

Elisabeth Stahl is the Chief Technical Strategist for the IBM Systems and Technology Group and has been working in systems performance for over 25 years.