When it comes to modernizing your data center, what’s good for the earth is also good for your business. Phasing in an environmentally-friendly data center can both lower your organization’s carbon footprint and reduce your operating costs. An enterprise with a large datacenter can save hundreds of thousands of dollars a year in energy bills by adopting sustainable and efficient IT technologies that minimize your environmental impact. This 12-part checklist can help you start the green IT transformation:
1. Review your organization’s IT infrastructure and inventory
The first step in any project where you propose to save costs, eliminate waste, optimize performance, and
increase efficiency is to know precisely what you already have. A clear inventory helps you understand what new equipment is necessary to purchase and what old equipment should be discarded. It also sets a baseline so that you can understand from what position you are making changes. This will help in performance, monitoring and determining the real ROI of the project.
This needs to be a two-part review. First, take a careful inventory of every server and its usage patterns, load, and power consumption in your data center. Second, make a similar list for all of the clients, switches, routers, andother power-consuming devices on your network. Include everything plugged in from wireless access points to storage devices to IP telephony and surveillance equipment to printers.
2. Review your physical facility and operations
Your IT operations are intimately connected to their physical environment. This means that a thorough evaluation of your current datacenter and future needs must include a facilities review. Can you consolidate multiple datacenters, for example? Or, can you disperse large power-sucking datacenters and create smaller, more nimble centers closer to the principal users?
You also need to evaluate the power supply and heating and cooling systems. The supply needs to be rightsized to avoid brownouts, and heating and cooling are intimately connected to the operational health of your IT equipment.
These last points lead to the need for an operations review. Does your IT operations team coordinate and
cooperate with your physical plant operations team and vice versa? If IT anticipates a surge in server traffic
and usage — maybe even a rapid deployment and provisioning of new servers — will they contact the physical operations team and make sure wiring, power and cooling requirements can be met?
3. Analyze IT operations need
With a thorough understanding of your existing enterprise equipment, you are ready to evaluate what new
network equipment you need. You should create a purchase plan for the next year, forecast usage for three
years, and even consider your organization’s needs five years down the road. This lets you avoid costly mistakes in failing to allow enough headroom for growth, which could lead you to have to rip out and replace equipment in as little as three years.
To make an accurate forecast, you need to survey business and IT stakeholders and ask them to project their IT needs a year, three years, and five years into the future. Your new, more efficient datacenter must be able to meet the needs of the business a year from now as well as the projected changes in three years.
Don’t forget to look at usage patterns, time-based usage, and peaks and lows. Even if you expect to make
substantial savings by consolidating servers into fewer datacenters, you have to allow for enough bandwidth,
processing power, and storage to handle peak network traffic.
Step 4: Analyze facility and physical operation needs
You need to make sure that the physical plant will operate effectively in lock step with your IT equipment, during both peak and low usage times. That means looking at locations, wiring, physical configuration, power supply, and heating and cooling systems. Make sure that you do not under or over provision any part of your operation.
5. Prioritize and optimize usage patterns
The next step is to optimize your operations. You need a clear understanding of which IT operations and business operations are mission-critical and have top priority. You also need to know how the different parts of your network are used to successfully handle these business applications, including their usage patterns. For instance, do certain business-critical applications run every night? Does traffic peak at a certain time every month?
Once you have a clear understanding of the expected operations and applications, their usage patterns and
their priority, you can optimize your IT practices to deliver them effectively. Ideally, you will have as much of
your network working optimally all the time as possible. In reality, you will need to be able to guarantee good performance for the peak utilization of IT resources without leaving too much of your equipment running idle during the lows.
6. Consolidate: Eliminate redundancies and unused equipment
Now, the real value of all the preceding work comes into play. Every single area where you can consolidate,
eliminate duplication, remove redundancy, and flat out get rid of unused equipment becomes an area where you make concrete bottom-line savings to your operations.
In organizations where employees have added to networks without supervision or where there have been large changes in organizational structure and operations, the savings at this stage can be substantial. Do not forget to document clearly the changes and to count power savings in your ROI as well as any capital equipment and ongoing operational cost savings.
7. Deploy virtualization technologies
At this point in the project you can start to take deploy server and storage virtualization technologies.
Virtualization is a powerful tool that can eliminate redundancies as well as unused bandwidth, storage, and
processing cycles by allowing one set of physical equipment to handle multiple workloads. This is particularly
valuable in an environment where usage levels move up and down continually.
8. Apply time- and usage-based provisioning
Modern servers can handle power cycling easily, and in hundreds of thousands of power cycles not one failure occurred during testing. This means that the efficient datacenter will use a WakeOnLAN-type technology for coping with peak usage. It also makes for considerable power savings. In addition, a WakeOnLAN technology combined with a smart provisioning manager and good virtualization means that the network can quickly adapt and adjust to rapidly changing loads.
These technologies illustrate why it is so important that a thorough priority usage analysis over time is made
for the datacenter. If that is done correctly, then this stage of the process can yield as many savings as the
consolidation stage.
9. Upgrade to energy efficient network equipment
Implementing a project with as much scope as a green datacenter, you cannot avoid equipment upgrades. It
is even possible that if no such analysis has been done for several years, a complete rip and replace may be
necessary. First, be sure to do a full audit and ROI analysis to make sure that a gradual change might not be a better plan. Datacenter, network and server technologies improve constantly. Some of the technologies you may wish to upgrade include blade servers, virtualization, dynamic provisioning, smart power monitoring, speedstep technologies that speed up and slow down processors to match demand and lower power use, WakeOnLAN, and others. Most modern network equipment is now designed to be energy efficient.
10. Working practices
This may be the single most important change introduced as part of the process. It is common for physical plant and IT departments to start out separately and then figure out ways to work together. However, for the modern datacenter you need to formalize the working relationship in some way. In addition, an operations group needs to also liaise with the regular business operations to make sure that communication is clear and swift about all forms of change — from outages to upgrades to communicating sudden business operation changes that could require significant datacenter support.
11. Monitor, monitor, monitor
As you phase in greener equipment, the evolving datacenter should be monitored extensively. Monitoring and
reporting allow you to accurately measure the effectiveness of the changes you’ve implemented. They also allow operations to do a significantly better job at running the new datacenter, which can in turn save considerable amounts in power consumption terms.
12. Recycle outdated equipment
This last item is an often neglected part of moving to a green datacenter — and even more often neglected are the cost advantages of doing this the right way. If old equipment still has some life left in it, you might be able to resell it or donate it and get a tax write-off. If it’s worthless, avoid sending it to the landfill by finding a local e-cycling facility or event. The U.S. Environmental Protection Agency can direct you to nearby resources.
https://s3.amazonaws.com/formcomposer/assets/asset/production/items/928/zd-12-steps-greener-datacenter_120711.pdf
