Important Data Center Metrics You Should Know

January 3, 2013 By

               We all know that data center and both mechanical and electronic equipment within the facility are becoming energy efficient. There are a variety of standards, certificates and metrics that have emerged to measure the efficiency of the facility. Metrics and Standards are very important to perform accurate measurement of data center and its IT equipment. Standards and Certification programs can maintain the integrity of IT companies’ claims for energy efficiency. Following are some of the certifications, standards and metrics that companies are using to show the energy efficiency of their facility and equipment.

Carbon Intensity per Unit Data:

               Carbon Intensity is used to refer the ratio of carbon-di-oxide to energy. It is a measurement of the amount of carbon emissions per unit of fuel consumed. Akamai is one of the leading content delivery network has been reporting carbon emission related to cloud computing. Carbon emission related to cloud computing based services in the metrics of “CO2/ megabytes of data transferred”. This metric enables Akamai to compare cloud computing usage across industry and also in process of making this metric to available on a monthly basis to its clients. Greenpeace has praised Akamai for its transparent reporting of this metrics and has awarded highest grade for its transparency above all giant internet players like Apple, Google, Facebook and, IBM.

Power Usage Effectiveness (PUE):

                This is the leading metrics which shows how efficiently the data center is operating. PUE is the ratio of how much energy is consumed in running data center’s equipment to the energy consumed to run the overall data center. This metric was created by “The Green Grid”. If a data center consumes lot of energy for cooling and power conversion, the facility isn’t very efficient and its PUE rating will be very near to 2. If very little energy is consumed in power conversion and cooling then data center will have a PUE closer to one.

Leadership in Energy and Environmental Design (LEED):

                This is a certificate awarded by the U.S. Green Building Council. LEED council evaluates green credentials of the building and consider several factors like materials, water used, indoor environmental quality and, location along with the efficiency of the plant to award any one of the five certificates certified by U.S. Green Building Council.

American Society of Heating, Refrigerating and Air- Conditioning Engineers (ASHARE):

                This society is focused on building systems, indoor air quality, and sustainability, refrigeration and energy efficiency. ASHARE also publishes a series of standards and guidelines relating to HVAC (Heating, Ventilation, and Air-Conditioning) issues and

systems.

Energy Star:

                Environmental Protection Agency (EPA)’s energy star certification is one of the widely used certification program for computers and gadgets. This certification can also be used for data centers and building that runs many number of data centers.

PAR4:

                This metric measures server powers in different methods including peak power, transaction per watt, idle power and total utilization of power.PAR4 metric is developed my startup power Assure and enables servers of different generations

to one another in terms of energy consumption and efficiency. PAR4 is gaining huge response from Internet giants like Cisco, Intel and, Dell and have been working to incorporate this metric into their systems.

 

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Data Center Management Tagged With: ,

Tips for securing your Big Data

December 10, 2012 By

               In IT industry Big Data is a collection of large data sets which are hard to manage on hand base management tools. Some of the challenges faced in managing big data are storage, curation, capture, analysis, visualization and search. Big data are difficult to manage using desktop packages, instead massive software running on several tens or thousands of servers. Demand for maintaining big data has increased especially in the field of telecommunication and IT. Some of the big data maintaining giants are IBM, HP, SAP, Oracle Corporation and Microsoft. Following are some of the tips which I came across for securing big data.

Plan about Securing Data before Starting Big Data Project

                Technicians should involve in protecting big data before installing or feeding data. Organizations need to find alternative ways to secure big data. Organizations should have serious discussions about data security and other related security issues.

Have A Clear Idea About What Data May Get Stored:

                Data incorporated into Hadoop for storing and running are subjected to specific security measures. Every organization faces huge risk or implementing security for data transfer and usage. So companies need to have a look at revenue loss and risk associated with leakage.

Centralize your account:

                Data are organized so vividly, sometimes data reside in diverse organizations. To ensure consistency in data access and for better policy enforcement, centralize the data account.

Separate Encrypted Data and Encryption Keys:

                Maintaining same server for storing both encrypted data and encryption keys is like locking the door and leaving the keys in lock socket. Key management system will allow users to store encryption keys and encrypted data separately.

Use Encryption for Both Data in Motion and At Rest:

                If companies plan to implement SSL encryption, this can protect data as it moves from applications and nodes. Encryption of data will protect malicious administrators or users to gain access to data. This is one of the cost effective ways to solve several data security threats.

Use Secure Automation:

                If you are dealing with the multiple node environments, this might hamper deployment consistency. So it’s better to use automation tools which help you to cheap cialis online configure the application, update the stack, certificates, patching and platform.

Use Kerberos Network Authentication Protocol:

                This is one of the effective methods for keeping rouge nodes and application of your cluster. This protocol can help to protect web console access. Kerberos is a painful method to set-up and re-validate new nodes and applications.

Add Log Files To your Cluster:

                Many companies started big data to maintain log files, why not to include log files on-to your clusters? Adding log files will provide you a place to look when you think you have been hacked or when something fails.

Ensure Secure Connectivity Between Nodes and Between Nodes and Applications:

                One of the best methods to ensure secure connectivity is by implementing SSL-TLS. If an organization implements an average built network then it is recommended to integrate the system with stack application.

                Do you agree to these steps, will these steps will help to secure big data in an organizational level.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Data Center Management

Data Center Operations and Its Biggest Mistakes

December 6, 2012 By

Data Center is a facility furnished with computer system and associated components, such as storage and telecommunication systems. Data centers are drawing attention globally with its unique features li

ke reliability, space for large data, redundancy and temperature control. Data centers generally include data communication connection, power supply backups, security devices, fire suppression and air conditioners. Massive data centers consume electricity as a small town and contribute a significant amount of air pollution in the form of diesel exhaust. Boom of data centers rose during the dot – com bubble. Similarly for years massive data centers are mesmerized to a lot of operational errors, mainly due to human blooper.

To yield better efficiency data centers companies have to ensure operational programs are fully functional and successful. Data centers have to be very watchful to avoid traps. Here are some of the common mistakes happen in data centers:

Relying Too Much on Data Center Design:

We can observe companies investing a lot of money on redundant design of the facility rather than concentrating on their operations. This is one of the best examples for human lackadaisical. Many data centers concentrate more on appointing specialist to design office building maintenance and no experts in managing critical operational facilities.

Disregarding Operation Team in Facility Design:

This is the first step that an organization needs to implement to balance the company’s capital expenditure. This step helps an organization to control and understand company’s operational expenditure. This step also includes determining performance characteristics and design criteria.

Failure in Addressing the Correct Number of Staffs:

Many Hosting companies owning data centers underestimate the need of appropriate staffing. Insufficient staffing can result in havoc at emergency situations. So it’s better to create schedule and deploy proper number of resources.

Failure in Training and Developing Right Talent

Once an organization hires a fresher or a qualified person, it’s important to provide training and support. Generating a positive

environment significantly improves the retention rates of the person. Companies need to have step by step training programs in order to expertise the work assigned to them.

Failing to Test Skills Consistently:

Organizations need to perform tests technicians consistently which might help in building right responses at the right time. Training alone cannot gain complete skills, regular test and assessments will improve overall reaction time during emergency situations.

Failure in Implementing Appropriate Procedures and Processes:

Companies have to look at the processes and procedures are moving on right schedule and yielding good results. Companies need to have a regular update on formal set of procedures and processes which return maximum efficiency. Some of the procedures include EOP (Emergency Operational Procedure), Emergency response, Standard Operational Procedure, vendor management and Method of procedure.

Failure in Using Software Management Tool:

Most of the data center organizations neglect to install software management tools to minimize the operational cost. This seems cost effective. When a company starts catering more number of customers, management team might be at risk because of poor document management.

Failure in Implementing Quality Systems:

Data center need to implement efficient, cost effective and reliable hardware. One of the best ways to ensure proper quality of the systems is by including a Quality Assurance team.

These are the main steps need to be maintained to gain maximum efficiency of the facility.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Data Center Management Tagged With: ,

Reasons to Renovate Your Data Centers

December 3, 2012 By

               Most of the data centers have following features false floors, windowless rooms, lots of electrical outlets, limited walking space and sterile. Data Center renovations one of the best paths where managers can improve performance and improve the overall efficiency of the center. By renovating Data centers, managers can yield a lot of improvements in terms of space, performance, operating cost and many others. When do managers know it’s time to renovate data centers?  These following points will give some idea.

               One of the main goals of data centers is to maintain cool temperature. Whenever you traverse inside the data center, if you get any hot flashes or when temperature fluctuates greatly. Then it’s time to check and renovate the cooling system. CFD (Computational Fluid Dynamics) is one of the gauges which tell how to manage cooling systems in your data centers.

               Each and every day technology keeps on upgrading.  Have a look at technologies running in the current market and implement those in order to gain more efficiency.  Implement new rack designs which provide a new level of savings in cost and energy and also improves the level of control over data center.

               No room to add more servers? When you think of adding a new rack of sever to gain more power into your data center, you think of renovation. Carefully analyze servers and remove the servers which are not running, this might cause under utilization of resources and might directly affect your operations and maintenance cost.

               A fan is one of the cheapest cooling agents in data centers. A fan is capable of redirecting hot air away from servers and allows cool air inside the servers; this can reduce hot air between servers the data centers.

               When your data center looks like a room of cables. This means data center needs renovation and management team has to improve cabling. This might money and time and this leads to overall improvement of the data center.

               Worrying about natural disasters? We are not sure when data

center might be struck with a flood or earth quake and other emergencies.  New data center designs concentrate more on disaster management and recovery, implementing any of these designs might help to overcome disasters.

               Managers need to have a close look at data center Infrastructure management tools and these approaches might be worth of looking into. DCIM tools favor small data centers which are scalable.

                Not getting expected efficiency from your data centers? If you think that you are not yielding expected efficiency from your data centers, then it’s time to renovate your data center. While renovating data centers look for alternative energy sources and alternative source for cooling your servers.  Managers need to make a regular survey on servers and eliminate redundant hardware which might eat up a lot of energy.

                Renovating a data center is far more complex than renovating a house. In data centers renovation is a continuous process. Renovation process is a critical issue which minimizes the risk of downtime and also enhances the overall efficiency of the plant.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Data Center Management Tagged With: ,

Is Modular Design for Data Center Right Choice for Future?

November 29, 2012 By

Modular design for data centers is a current hot topic in IT industry. Companies are struggling to meet IT need and maintain efficiency. If data center companies choose to include modular design, then they have to select pod or containerized design. Modular design will provide flexibility to data centers. Some of the benefits enjoyed by modularized data centers are:

 Speed:

Modular Data Centers can be built 75% faster than conventional data centers.

Efficiency:

In Modular Data Centers hardware can be addressed incrementally, rather than immediate constructing entire data center space. In Dell data center solution engineering team are working hard to optimize entire data centers to reduce operational cost.

Flexibility:

Modular Data Centers are designed with standardized infrastructure which enables data centers to use operating system universally. This design will allow designers for easy replacement of hardware and operating system. A truly modular data center is either incrementally constructed or uses containerized solution.

 

Altering a traditional data center to Modula Data Center is similar to constructing a new facility with blocks of buildings, so it’s preferred to implement containerized approach in such conditions. Now let us have a close look about POD design of data centers and Containerize solution

Pod Design for Data Centers:

In pod design data center consist of IT room, a supporting infrastructure room and IT computer room.  Supporting Infrastructure room includes mechanical and electrical room which provide power, cooling to IT computer room. Later design engineers can include any number of IT computer room based on their load requirements.

Instead of supplying space, power and cooling for an over build data centers at 50% efficiency rate, companies can divide data centers into small pods. Let us have an example, suppose a company with 10,000 sq ft area, companies can estimate a facility based on requirement and build a small pod of 1000 sq ft of storage and server equipment. Company can include adjutant pods based on IT requirements by following “pay as you grow” approach. This approach can lower the overall operating expenses and also increase the efficiency of the data center.

Containerized Solution for Data Centers:

Instead of designing a pod design from scratch companies can choose containerized solution, if company is ready to build, relocate or expand its facility. By following this containerized solution companies can lower their operating and capital cost, companies can also be benefited by avoiding local building techniques and codes. Cooing systems and airflow to servers can be engineered before implementing in the plant to optimize efficiency and performance. Containerized data centers can cut down deployment time from one year to weeks. Containerized data centers are best suited for remote locations. Containerized solutions gain more points in terms of mobility, containers can be placed anywhere and all containers will be pretested before deploying to the centers. Containerized solution can also be used as temporary solution when companies have expansion plan or when data centers come across any disaster. Apart from all these Containerized solution will provide multi-location support and high redeployment capabilities.

Before selecting any of the two modular design, it’s better to have a look at advantages and disadvantages of the two designs.

 

Pod Design

Containerized Solution

Flexible in size

Equipments must go in pre-fabricated containers

Canbe modified based on requirements

Newly configured containers must be deployed

Space can be designed for comfortable services

Uncomfortable and offers limited space for technicians

With proper design maximum efficiency can be achieved

Designers need to spend significant amount of dollars to improve efficiency

Permanent solution

Best suited for temporary needs

Initially takes at least 6 months later inclusion can happen more rapidly

Can be deployed within weeks

Must follow local building codes

Must follow ISO complaints

 

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Management Tagged With: , ,

Ways to Reduce Water Consumption in Data Center

November 27, 2012 By

All data center management teams have been thinking about conserving renewable energies, but recently data centers are concentrating more on conserving water. Saving any amount of water saves money, because water in short as it’s mostly used in domestic purposes. Data centers require an enormous amount of water to cool cloud computing server farms. James Hamilton Data Center Designer and Researcher in Amazon estimates Data centers with a capacity of 15 Mega Watt can utilize up-to 360, 000 gallons of water per day. Hamilton added, “Water is tomorrow’s big problem, no one talks about water. The water consumption (in data centers) is super embarrassing. It just doesn’t feel responsible. We need designs that stop using water.”

          Google and Microsoft are adopting new alternative methods to cool their massive data centers. Companies are using water from nearby rivers canals and also using recycled water. Microsoft recently added a new Data center in San Antonio because local water company can provide large quantity of recycled water; this would have less impact on city’s drinking water. In Belgium Google installed a new data center next to Industrial canal, data center can draw water to cool massive servers using industrial water and can draw water from wells and rain water. Google has designed a new cooling system and they have named it as “Hot Huts”. These Hot Huts are temporary homes for hot air which leaves main servers. Fans are installed on these huts will pull hot air from behind the servers and lead them to water cooled coil, these ambient hot air from the servers. These fans will help to cool the air by leading them to water cooled coils and thus completing the cooling cycle.

One of the best methods to conserve water is by re-cycling it. Google’s data center in Hamina, Finland uses sea water to cool servers without using any chillers. This data center is located in Gulf of Finland. Cooling systems pump cold water from sea and heat from servers are exchanged. This ambient hot water is cooled before returning to gulf. Another Google’s data center in Douglas County uses city waste water to cool its servers. In this data center large tanks are filled with this water and fin sand particles are filtered out and leaving clear water to cool its servers (this water is unsuitable for drinking purpose).

Data centers must maintain their cooling systems regularly to avoid excess consumption of water. All Data centers must implement such alternative methods to cool its massive servers to reduce water consumption.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Cooling, Data Center Management Tagged With: ,

Which is More Reliable in Data Centers: SSDs VS HDDs?

November 27, 2012 By

Data Centers mainly rely on storing or accessing multimedia contents. Data centers mostly struggle with the problem of how to store information; companies can opt for any of the technologies like Hard Disk Drives (HDDs) or Solid State Drives (SSDs). But which technology is the best, HDDs or SSDs?  In current market situations, solid state drives are more reliable than hard drives, but this isn’t always true in all conditions. Before getting into any conclusion, let us know what are SSDs and HDDs.

SSDs and HDDs are storage devices, but have different storage methods. HDDs store data magnetically on spinning platters and use a mechanical read or write methods.  On the other hand SSDs store data electronically in semiconductor circuits in binary (1s & 0s) format. SSDs are employing nonvolatile flash memory as their main storage medium, nonvolatile medium stores data even when the power is down.  Differences in their storing technologies lead to number of advantages and disadvantages. So before replacing all HDDs to SSDs have considerable idea about both the technologies.

Google has made a detailed study on why HDDs fail. Following are some of the conclusions drawn from the study, 36% of failed hard drives did not exhibit any SMART (Self Monitoring, Analysis and Reporting Technology) monitored failure, because failure of HDDs might have been caused because of mechanical or electrical failures. HDDs are mechanical devices whose components would wear out due to lose precision or dry lubricants, as a result they draw greater amount of power and emit greater amount of heat. SMART system is almost useless because I cannot detect mechanical failures modes which include capacitors, power regulators, connectors and firmware.

SSDs replace mechanical components of HDDS like heads, motors, platters and bearings with flash, but SSDs don’t replace any electrical components that cause many failures. SSD chips are manufactured in different grades, manufactures like Intel Samsung and Toshiba manufacture highest graded chips for their own SSDs and lesser quality chips are sold to market and most of such lesser quality chips go into SD cards and USB drives. Lower quality chips have more number of failures compared to higher quality chips. One more advantage of SSDs are its writing and reading speeds, SSDs can read or write data at a rate of 200-500MB/s, whereas HDDs reading and writing speed is within 50-120 MB/s. In high performance environment where fast booting is needed, this high speed read and write capability is more advantageous. Another advantage of SSDs when compared to HDDs are its power consumption, SSDs consume almost 50% less power than HDDs. In addition to all these disadvantages HDD produces more vibrations and noise, and SSD can withstand more physical damage than HDDs.

Pricing of SSDs are more when compared to HDDs, storage capacity of HDDs are more when compared to SSDs. SSDs offer numerous advantages over HDDs, but it cannot compete in terms of mass storage capacity and cost.

Hybrid storage systems can be implemented to overcome the disadvantages of each technology. Data centers can include expensive SSDs where faster read and write or booting is required and it can include inexpensive HDDs where large storage is needed.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Management, Data Center Solutions Tagged With: ,

Data Center Efficiencies from the Perspective of the End User

November 22, 2012 By

Today, data centers have become the prime targets of environmental agencies because of their supposed incessant use of energy. Hence the need of the hour is to focus more on consolidating data centers and increasing energy efficiency. After controversies surrounding some of the bigwigs of the Internet arena for not being responsible and considerate towards the environment, data centers’ attention mainly revolves around optimizing operations and efficiency. Several parameters define an efficient data center.

Some of the highly efficient data centers include infrastructure such as advanced cooling techniques, high density server racks, rooftop cabling systems, biodegradable energy utilization techniques along with minimum utilization of space. Data center operators strive to implement radical measures to further enhance the efficiency of a data center. The efficiency can be further enhanced by deploying rack optimization techniques, monitoring tools, and virtualization strategies.

Analysis has revealed that there is often an overlooked factor that can apparently take the efficiency of data centers to another level. The factor is “looking at data centers from the end users point of view.”

Relation between a data center and the end user

It is a known fact that any new product or service is delivered keeping in mind the end user and end user satisfaction. Then why leave the end user in darkness when considering the efficiency of data centers? Data center efficiency when built amidst end user’s computing experience creates a more conscientious environment. This is where the idea of ‘user virtualization’ is used. User virtualization is the process of isolating/disconnecting the user from the hardware. With this users can carry all the settings, applications, and profiles regardless of the devices, hardware or even the operating system.

Hence with this new strategy users can have all the applications with them at all times, on any device, and on almost any platform. But what needs to be answered here is that if user virtualization can really create data center and IT administrative efficiency?

BYOD and IT consolidation: BYOD stands for ‘Bring Your Own Device’, and as the name suggests it can reduce the number of resources required within the data center and the IT environment. IT consolidation can further reduce management issues, end to end physical points along with a simplified computing process. User virtualization allows users to carry data from one device to another seamlessly without having to depend on the hardware. All that matters is the end client receiving the data. All of this further translates to fewer moving parts within the IT environment.

Data Center Server Reduction: As most of the servers catering to BYOD are Virtual Machines, there is considerable reduction in the number of physical servers provisioned at the data center level. High density computing can further reduce stand alone rack servers, and create a more efficient environment.

Cloud Computing: With the help of user virtualization it is easy to transfer data to the cloud platform. User settings can easily centralize and controlled at the data center level or within the cloud environment. All settings can be controlled from one management console. User management functions can also be outsourced to the cloud. Hence, resulting in the reduction of hardware platform and software management tools.

By creating a more robust end-user experience in deploying user virtualization, data center administrators can find themselves working with a new layer of environment efficiencies.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Management, Data Center Solutions Tagged With: ,

Data Center Infrastructure Management

November 16, 2012 By

 

Most organizations spend more than 70 percent of their IT budgets on infrastructure such as servers, operating systems, storage and networking. This is required to manage the swift yet uneven pace of growth that is essential to survive in today’s highly competitive environment. When organizations achieve the critical mass, they are often bogged down by the widely varying demands placed on their IT infrastructure.

Even as the investments for hardware and software are tracking downward, the costs of managing and supporting infrastructure are increasing. Server sprawl leads to increased operations and support costs, labor intensive legacy platforms lead to increased administration and support costs, deployment and maintenance of IT systems is costly and consumes too much time and effort and, managing system security and user identity information across enterprise systems can be complex.

Desktop, Server and Device Management

Data centers can manage desktops, mobile devices, and servers as well as deploy patches, operating systems, and applications across the network, asset inventory and proactive monitoring of systems by implementing and managing the Microsoft System Centre Suite.

System Center Configuration Manager (SCCM) – The SCCM comprehensively assesses, deploys, and updates servers, client computers, and devices-across physical, virtual, distributed, and mobile environments.

System Center Operations Manager (SCOM) – SCOM uniquely enables data centers to reduce the cost of data center management across server operating systems and hypervisors through a single, familiar and easy to use interface.

Security and Networking

Data centers should guarantee that information and communication are protected from unauthorized access while at the same time provide a mechanism to protect the IT infrastructure from attacks and viruses while preserving access to corporate resources.

Fore Front Endpoint Protection (Anti-Virus) – By allowing a data center to use its existing client management infrastructure, the solution helps reduce ownership costs while providing improved visibility and control over endpoint management and security.

Fore Front Threat Management Gateway (Firewall and Proxy) – The solution reduces the cost and complexity of web security through multiple layers of continuously updated protections that are integrated into a unified, easy to manage gateway.

Fore Front Unified Access Gateway (Firewall, Proxy and SSL VPN) – The solution utilizes a combination of connectivity options, ranging from SSL VPN to Direct Access, as well as built in configurations and policies. This enables the delivery of comprehensive, secure remote access to corporate resources for employees, partners, and vendors on both managed and unmanaged PCs and mobile devices.

Data Protection and Recovery

Organizations require a structured or disciplined backup, storage and restore management solution to protect their data. Disaster Recover Planning – an in-depth analysis of current systems and creation of definitive plans in case of natural or intended disasters.

Virtualization

In today’s highly automated, technology-enabled economy, business capability has become inextricably linked to IT capability. However, IT infrastructure is becoming increasingly complex. The need of the hour is to leverage IT so that it becomes a more efficient and strategic asset of business.  By consolidating resources through server, client and storage virtualization, organizations can increase their resource utilization from 10%-15% to as much as 80%. Hardware requirements can also be reduced by 10:1 ratio or better. Automated system management and administration at every level in the data center allows greater focus on more strategic planning activities. Moreover, through virtualization, organizations can achieve a reduced deployment time for new IT services by 50% to 70%.

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Management, Virtualization Tagged With: , ,

Data center efficiency-Is it size dependent?

November 9, 2012 By

Data centers for a long time have taken their own stand when questioned about the efficiency of operations. Some data centers follow the conventional methods of strategy implementation while the others take a different route. For example, Google has adopted custom built servers and batteries, Facebook uses a unique evaporative cooling method to save power, and Yahoo has adopted an outdoor cooling design.

Data centers can be generally divided into two primary classes based on size that is the large data centers and the small data centers. Large data centers have an inherent advantage in implementing energy efficient strategies as they can shift operations to another facility as and when the situation demands. It could be a failure in the data center or an expansion plan adhering to all energy saving norms.

“Large centers can operate at a higher risk of failure with the offset, but smaller data centers can’t do that as easily,” said Fluegeman, who specializes in facilities, power and cooling, and is also the principal engineer with PlanNet Consulting.

“It’s following the sun or the moon, because they can very easily shift their traffic to different locations depending on whether it’s night or day in a particular location,” he said. “Especially if you’re using outside air and free cooling or air-side economy as it’s called, this is a concept that’s used to some degree by larger companies with data centers worldwide, where they shift traffic.”

In the case of smaller data centers (data centers that lack the infrastructure and resources of its larger counterparts) energy efficiency strategy implementation along with reducing costs is a difficult task.

So how does one cope with such a scenario?

According to Fluegeman, three quintessential parameters to be followed by a small data center are-

Design of the IT structure: it is essential to design the data center in a way that it helps in consolidation of the files. For example using file duplication or compression. This would automatically reduce the amount of space required for the storage of files. Enterprises can modify or make alterations to the network configurations from their own data centers as well as allowing the users to access data from anywhere at any time.

“There are ways to architect your system to deal with duplication and keep things lean and mean and trim with good housekeeping,” he said. “This allows data centers to be smaller and greener by using as little power as possible.”

Tiered storage can be used as another remedial measure: with this files are stored in servers, and the servers run on sleep mode. The files can be pulled up or retrieved only when necessary, thus saving a lot of power.

Restructuring the cooling method: It is imperative that small scale enterprise data centers make use of hot and cold aisle containment for data center cooling. This is to ensure that hot air is isolated from the cold air. But the crucial aspect of this design is to keep the hot air from nixing with the cold air. This would create a 30 degree difference in the temperature of the data center, says Fleugeman.

One way is to line up the servers such that the front faces one direction and the exhaust side faces the back. This forms the hot aisle and the cold aisle. Another way of doing this, in case of parallel arrangement of servers is to stack servers with alternate hot and cold aisles. Front faces of two servers face each other while the back faces the front of another line of servers

Use uninterrupted power supply to reduce excessive power costs: The use of the UPS is a radical transformation in the deployment of energy efficient measures. UPS is normally used when there is a power outage, but this can be prove to be an advantage for data centers as UPS wakes up from the sleep mode only when the power fails.

The difference in power drop and the UPS turning back on is so minimal that the equipments fail to pick up the drop in power and continue to operate continuously.

Switching to higher voltage equipments can result in a lot of power savings since it is more efficient and the physical conductors are smaller.

“If you look at power lines on big poles on the hill side, those wires are small- they’re far apart, but the power voltage is very high,” he said. “This is being applied to data centers now. Higher voltage is slightly more dangerous and in the U.S we have a lot more attorneys. That’s what keeps voltage lower, in my personal opinion.”

 

Data Center Talk updates its resources everyday. Visit us to know of the latest technology and standards from the data center world.

Please leave your views and comments on DCT Forum

Share on TwitterSubmit to StumbleUpon
Filed Under: Articles, Data Center Cooling, Data Center Management, Data Center Power, Data Center Solutions Tagged With: ,
« Older Posts