IT Agility: better use of energy monitoring data

Designed for the Internet of things, today's data center hardware devices provide valuable feedback, making it possible to automate full software definition.

Jefkraus

Every year it needs to do more, including managing an increasing number of servers, smart devices, applications, and services. End users are constantly raising standards, demanding higher performance and better response times. At the same time, everything in the modern data center continues to evolve, while the number of employees and operating expenses budgets remain constant or even reduced. Data center managers and their teams must constantly find smarter ways to work and reduce the maximum cost of spending. As energy costs have become a top priority, finding ways to effectively reduce power consumption is a good start.

The widespread use of virtualization and cloud technologies helps IT departments to build and manage data centers more cost-effectively. However, since all of the focus is on the high-level data center model, it is easy to overlook some of the significant advances made in the hardware-level area. The latest servers, storage devices, switches, racks, power distribution components, cooling devices, air handlers, and many other component devices provide timely feedback of status information to the network, especially as the server has evolved to offer a wide range of up-to-date and granular monitoring and control options for remote IT teams.

Information on practicality

At first glance, network-based built-in monitoring and control capabilities are appealing, but it teams cannot constantly query and tweak a single device. It is more impractical to manually collect enough data points to identify the appropriate patterns and trends. The solution provider continually improves the system console and data center dashboards to automatically collect, summarize, report, and document a wide range of device status information using middleware technology.

At present, most large data centers in the world use a full Software data center architecture. Highly automated IT operations include monitoring the server's real-time ingress temperature and power consumption data sources from rack/blade servers, Protocol data unit PDUs, and UPS. In addition, the airflow is monitored.

The industry's best energy management solution digests this information and translates it into energy and thermal images in every room and data center. At the same time, the aggregated intelligence data can help data centers and facilities teams better understand and manage energy costs and make better decisions by effectively combining control capabilities such as power capping and dynamic server frequency tuning.

For example, the energy consumption data collected in real-time enables a significant improvement in capacity planning. In the past, IT departments had to rely on the manufacturer's peak power specifications to estimate power supply specifications, or to use simulated loads to measure them. As more power consumption data is collected from production servers, IT departments are now able to more confidently and proactively configure new servers and racks. Power capping capabilities, another feature that modern servers can use today, can be used to ensure that denser racks do not exceed the maximum power limit. With the success of avoiding the risk of IT equipment power-down, higher rack density can fundamentally reduce the number of racks required, as well as floor space and cooling capacity.

650) this.width=650; "src=" Http://s1.51cto.com/wyfs02/M02/7F/6A/wKiom1cdxruwJzOvAAHWQ4OTxqs929.png "title=" It agility-web_1.png "alt=" Wkiom1cdxruwjzovaahwq4otxqs929.png "/>

In addition to providing more accurate capacity planning, the effective power, temperature, and airflow status support can be further improved:

• better data center operations. in addition to identifying hotspots, thermal images can indicate over-cooled channels. The IT management team can also take advantage of the increased visibility of power and temperature changes to adjust the ambient temperature in the server room. From the point of view of cooling costs, the increase in data centers per degree of Celsius can bring significant cost savings each year, so many data centers gradually accept higher room temperature levels, especially advanced servers and equipment can withstand higher temperatures.
  

• asset utilization and integration. The power consumption model can effectively find "ghost" servers as well as idle and underutilized servers. The server uses about 50% of its maximum power rating even when it is idle, so an effective assessment of whether the server is fully utilized can bring significant savings. It teams can consolidate servers or use dynamic real-time tuning to adjust idle servers to more power-saving sleep patterns.

• effective planning of workloads. work planning can be effectively implemented even in highly virtualized environments, taking into account the impact of overall power consumption. The migration and assignment of energy-aware virtual machines provides support for more energy-efficient operations, and increases awareness of energy costs related to specific tasks or organizational workloads.

• Equipment Life Optimization. real-time thermal images effectively highlight hotspots and proactively avoid damage to the most load-carrying and mission-critical servers. Helping to protect the power-capping function of the rack is also helpful in easing the heating problems of the equipment. Similar heat problems can cause damage or shorten the life of the server and other data center devices. Or IT teams adjust cooling and airflow systems to solve and eliminate hot issues, or reduce workloads.

• continuity of business. better understanding of the actual power requirements of different services, different systems, or different user groups, it teams can effectively adjust the disaster recovery plan to more intelligently allocate backup resources or transfer workloads during power outages. Based on the practical experience of operational operators in multiple data centers, more intelligent distribution of power usage can effectively extend the backup power life of up to 25%.

• SLA Management. IT management can set up a power policy to ensure the best execution of critical priority services. The automatic threshold management system, when the system, rack, or unit is approaching the limit, alerts the IT system to proactively adjust resources before the limit values negatively impact service levels.

• avoid high electricity bills at peak times. Many large companies distribute data centers geographically to provide the best possible service for each location. With the development of the visualization technology of power consumption model, IT system can reduce the cost by dispatching some workloads remotely and effectively using the non-peak electricity charges.

Higher controllability, lower operating costs

Data center operations teams use automation and middleware technology to gain more agility and control over data center resources. IT systems and facilities can better align and distribute data center assets while reducing energy costs in the data center.

At a high level, software systems for energy consumption and temperature detection are also used to assist in the adjustment of energy management policies for servers, racks, units, computer rooms and the entire data center. This information contributes to the green innovation and development of data centers and the formation of energy-saving effects. In addition, some data centres use power data to launch more accurate fee-based services. Based on published findings, the Intelligent Energy management solution can produce an astonishing 20%–40% reduction in operating costs by reducing energy consumption alone.

It is important to note that the increase in agility and cost reduction is not just a result of monitoring. As mentioned earlier, the latest generation of data center devices makes it possible to adjust the key parameters of a device over a long distance. For example, the IT department can dynamically adjust the internal power state of the data center server and the frequency of the processor's operation.

The software system provides the IT infrastructure team with visual and intuitive data aggregation of environmental conditions by providing energy and temperature data to the console and dashboard. The aggregated information actually improves collaboration between IT systems and infrastructure teams to better coordinate the planning and management of the IT infrastructure.

The perfect combination of data monitoring and controllability mentioned above enables it to strike the best balance between server performance and power consumption without a noticeable drop in user experience and service levels. Field tests show that the dynamic adjustment can reduce the power consumption of the server by about 20%.

Conclusion

Global data consumption has risen rapidly in our highly connected world, resulting in an upward trend in data center energy consumption levels. The Natural Resources Conservation Commission NRDC reports that there are 10% (91BKWH) of global energy use being generated by global data centers. Power and cooling costs have become the largest component of the data center operating budget. Therefore, for any data center, how to obtain the actual power consumption of the data center image display and thermal map is considered to be the highest priority target.

Fortunately, IT operations teams can assemble and apply real-time data that already exists throughout the data center through highly automated monitoring and control solutions. By avoiding the energy waste and the significant savings resulting from an effective extension of equipment life, the investment in Software Solutions provides an effective business case, accelerating deployment and start-up time.

The industry's best solution now eliminates the need to install any client hardware and software on the server. Easy-to-integrate software devices reduce the burden on IT staff and provide more effective tools for IT infrastructure and facilities staff.

650) this.width=650; "src=" Http://s2.51cto.com/wyfs02/M02/7F/68/wKioL1cdx6exdLBtAAH0D9bx-X8817.png "title=" It agility-web_2.png "alt=" Wkiol1cdx6exdlbtaah0d9bx-x8817.png "/>

The latest generation of overall energy management solutions represents a major step forward in data center monitoring and management systems, and the middleware approach has proven to provide the necessary scalability to maintain synchronization with the data center. An open and flexible software architecture is also in line with today's flexible, on-demand service delivery model. Hardware vendors and system integrators look for more scalable feature settings and deployment options with evolving data center smart hardware and interconnect standards.

For more product information, please log in:

Http://www.intel.cn/content/www/cn/zh/software/data-center-overview.html,

or send an email to [email protected] to contact us.

Author Introduction:

Mr. Jefkraus is the general manager of Data Center Manager (DCM) solutions for Intel Corporation in the United States. At Intel Corporation, Jefkraus has more than 13 years of team management experience. His team is developing a middleware product for energy and thermal management. Mr. Jefkraus holds an MBA degree from Boston University in Business administration.

This article is from the Intel Data Center Management specialist blog, so be sure to keep this source http://inteldcm.blog.51cto.com/11451124/1767542

IT Agility: better use of energy monitoring data