Data center location considerations ,infrastructure and management

The choice of where to locate your data center is one of the most important decisions that you will ever make as an IT Infrastructure specialist. Making the right choice will ensure that your business can continue to function and meet customer expectations while making the wrong decision could be catastrophic. This article seeks to help you choose the best possible location for your data center, by examining five distinct locations for construction: Cold, Warm, Mixed (Warm), Tropical (Warm) and Hot (Tropical).

Data Center Location – Cold

Cold climates are typically characterized by large snowfalls during winter months, with shorter-lived snowfalls occurring in some areas that experience summertime temperatures that average less than 12° Celsius. Data centers located in such climates must consider electrical infrastructure limitations, as well as the effects of freezing conditions on personnel who are required to gain physical access to data center equipment.

One example of a cold climate data center is located in Sainte-Thérèse, Canada. Average winter temperatures are typically below -15° Celcius and can reach lows of -40°C during some extreme cases. To deal with these low temperatures, the facility has implemented numerous heating methods including non-mechanical heaters, auxiliary boiler systems, and hot water pipes. These heating methods have contributed to an operating cost per kW/h of 7 cents for this location.

Data Center Location – Warm

Warm climates are typically characterized by moderate temperatures throughout the year, with little or no variation between summer and winter. However, certain climate conditions can lead to extremely high levels of humidity and precipitation over extended periods, which must be accounted for in a data center design. For example, heat generated from IT equipment inside a facility could contribute to increased humidity by indirectly warping server components through condensation. Data centers located in such climates should consider installing roofing solutions that have been specifically designed to drain water away from the exterior walls of a building as well as using air conditioning systems that provide adequate cooling without exacerbating humidity issues.

The Interxion campus in Paris is one example of a warm climate data center location. Located in the Paris metropolitan area, this location experiences an average temperature of 13°C during summer months and 4°C during winter. The site’s cooling systems are optimized to work within these limits, with multiple raised floor spaces that allow cold air to enter the facility at low level, while heated exhaust air is removed through roof vents. This design allows Interxion to achieve a PUE of 1.1 for their campus, with an annual energy cost per kW/h of 6 cents.

Data Center Location – Mixed (Warm)

Mixed-climate data centers are typically located in regions where there is significant variation between seasons throughout the year. For example, North Carolina has a mixed climate because it experiences significant variations in weather conditions between summer and winter. This includes higher than average rainfall during the summer months when compared to other U.S states.

The TelecityGroup campus in Newcastle is an example of a mixed-climate data center location. Located in the North East coast of England, this site experiences mild summers with an average temperature of 15°C and very cold winters with an average temperature of 6°C (source: http://www.timeforyourdata.eu/europe-24/uk/newcastle#loc=34). The facility’s cooling systems are designed to provide adequate cooling without humidity issues, which has allowed the site to achieve a PUE rating of 1.13 with an annual energy cost per kW/h of 7 cents.

Data Center Site Selection Considerations

While there are many factors that go into selecting the ideal data center location, these five climate variables were chosen for this article because they represent some of the most significant challenges faced by data center operators worldwide today. It is important to note however, that site selection decisions must also be made based on power availability, land availability, connection speed options and Internet capacity constraints. Finally, it is essential that any potential facility operator conduct a thorough due diligence process involving current market conditions as well as forecasts for future economic growth at each prospective site before making a final decision. Only with adequate time dedicated to site analysis will an organization uncover all of the opportunities and pitfalls associated with bringing their next-gen data center online

Datacenter infrastructure management

(DCIM) is a hot topic in big data analytics nowadays. DCIM software allowed users to monitor, manage, and optimize various resources within the hardware layer of ICT infrastructure. These resources are usually categorized into two layers: the electrical layer which includes power usage effectiveness (PUE), the total cost of ownership (TCO), etc.; and the thermal layer that deals mainly with the cooling system.

The most important part of DCIM is energy management which deals with how effectively the energy is used by every component in a computer room. The efficient use of energy can reduce operational costs, improve the Availability of Infrastructure for Applications (AIA), help prevent service interruption, increase throughput while maintaining high levels of reliability, boost underutilization of facilities by identifying inappropriate workloads and balancing them across server farms, and so on.

DCIM enables IT, managers, to monitor power draw over time in order to determine the current and real-time energy usage within a data center or computing environment. However, existing DCIM systems still suffer from some major limitations: they do not address the correlation between power consumption and performance metrics such as response time, throughput, utilization of servers/resources, etc.; foster management information overload; cannot provide a diagnosis for complex problems such as finding congestion points in computing clusters or track issues at very granular levels; lack of users’ ability to drill down into every detail which makes it hard to correlate with various business operations; and established tools fail to leverage machine learning algorithms.

At this point, we’d like to introduce our Data Center Infrastructure Management Software (DCIM) where we tried hard to eliminate those shortcomings and leverage the machine learning technology as much as possible. DCIMS is open-source software that provides power management and thermal monitoring for big data clusters. It allows you to monitor servers/resources within a cluster so you can identify resource over-utilization and wasted energy. You can also manage your power supply more effectively through our web console or RESTful APIs by switching designated servers on and off manually and automatically based on business requests such as hypervisor call, user requests, etc.

Our main features include:

monitoring server energy usage detected by embedded sensors; powering down/up servers to cut down energy waste; online power supply switchover to avoid service interruption; pass-through cooling using hot/cold aisle arrangement which can reduce servers’ power consumption by 20%–50%; real-time CPU utilization and system load monitoring to help you balance your compute cluster’s workloads more efficiently.

The DCIMS software is designed for big data center operators, server room managers, etc. It can be a valuable asset for IT professionals who are concerned about the growing cost of operations and reducing carbon footprint in their organization. You can download it at (no registration required) or check out our GitHub repository . We welcome any comments or suggestions as this project is still under early development phase. Your feedback will help us make DCIMS better for everyone!

Please note that currently, DCIMS supports only CentOS 7. The other Linux distributions will be supported in future versions. Stay tuned to learn more about us and our latest updates by following us on Twitter , Facebook , LinkedIn, Google+, YouTube.

Datacenter virtualization

Datacenter virtualization is the new buzz word and VMware is at the forefront of this market space. According to IDC, “Server Virtualization software revenue grew 44% in 2009 while Server Hardware shipments declined 12%. This is indicative that IT organizations are increasingly doing more with their existing servers through virtualization.”

According to Gartner, “Virtualization technology will be one of the top strategic technology trends for 2010 and beyond.” They also predict that by 2013, 50 percent of all x86 server workloads will execute completely in virtual machines or other abstracted form factors such as Cloud Computing environments. [Source: Gartner’s Steve Prentice presentation on Data Center Decisions Day – October 16th.]

VMware has a good hold over the data center

VMware has a good hold in the virtualization space and is expected to be at the forefront of this market. With an increasing number of applications and OSs, VMware has also increased its product portfolio including:

Datacenter virtualization

is helping IT organizations reduce IT management costs by consolidating multiple workloads into a single physical server or into a cluster of shared servers. Applications that were earlier running on separate hardware can now run together on one hardware with low cost per application. This approach also helps companies avoid costly system upgrades resulting in optimum utilization of existing infrastructure investments. Application owners can automate their processes and improve service levels by creating test/dev environments for showcasing new products before deploying them in production reducing risk and delivering better business outcomes.

Advantages of Datacenter virtualization

Cost Savings: Reduce IT management costs by consolidating multiple workloads within a single physical server. Avoid costly system upgrades resulting in optimum utilization of existing infrastructure investments;

Improved agility: Automate processes and improve service levels by creating test/dev environments for showcasing new products before deploying them into production;

Increased business value: With reduced risk, organizations can deliver better business outcomes.

Disadvantages of Datacenter virtualization

The existing hosting model needs to evolve;

Not suitable for the software that requires direct hardware access (physical-style). Licenses may not be transferable across the shared servers.

New licensing models are needed to cater to data center virtualization. Virtualization is being used to reduce IT management costs by consolidating multiple workloads into a single physical server or into a cluster of shared servers. Applications that were earlier running on separate hardware can now run together on one hardware with low cost per application.

This approach also helps companies avoid costly system upgrades resulting in optimum utilization of existing infrastructure investments. Application owners can automate their processes and improve service levels by creating test/dev environments for showcasing new products before deploying them in production reducing risk and delivering better business outcomes.

Disadvantages of data center virtualization are the Existing hosting model needs to evolve, Not suitable for the software that requires direct hardware access (physical style). Licenses may not be transferred across the shared servers. New licensing models are needed to cater to data center virtualization.

Why VMware?

• Strong partner ecosystem: VMware continues to expand its customer and partner base by offering best-in-class solutions for the complete virtualization market. Customers benefit from a strong ecosystem of technology providers who work with VMware – including storage, networking, and systems management partners – all delivering exceptional results and expertise in their respective areas of focus. Plus, new value-added resellers (VARs) can quickly establish themselves as leaders among their customers by harnessing these technologies to meet those customers’ needs;

• Groundbreaking products: VMware delivers unique industry innovation with every release of our products, creating superior business value for our customers through an open, extensible platform

Leave a Comment