Green IT Innovation – Data Center Efficiencies to Reduce Carbon Footprint – Part 3 of 3.

The Green Data Center

Greening Up your Data Center from the Inside Out.

Part 1 and 2 of this series dealt with server consolidation and access to green energy as ways to help reduce your company’s carbon footprint.  This article gives you front-line knowledge about Data Center energy efficiency changes that really work and don’t require large capital investment.

Every DC is different, so we’ll focus on general things that should work no matter how your DC is configured.  One of the best places to start is getting a LEED assessment of your building(s).  The Leadership in Energy and Environmental Design (LEED) Green Building Rating System™ is the nationally accepted benchmark for the design, construction, and operation of high performance green buildings. This assessment will provide you with a baseline report with current and potential scores based on criteria like whether your air conditioning has an economizer mode.

The bank where I worked had an assessment done for our DCs and it identified several quick and inexpensive changes:

1. Monitor computer room humidity and program all AC units to operate in a way that prevents them from “fighting” each other (one unit humidifies while another dehumidifies).
2. Measure the cycling of AC units within your cross connect and high density rooms so that only the units required for cooling the load are operating, leaving the remaining units in a standby mode. Be sure to switch “lead” and “backup” AC units on a regular basis to avoid over-stressing one unit.
3. Conduct a raised floor survey to verify floor cut-outs and penetrations are plugged and hot-cold aisles are configured correctly; resolve as necessary. This will need to be monitored and changed as your DC configuration morphs over time.
4. Check your server operating specs; you can often increase the raised floor ambient temperature and broaden your humidity set points.
5. Replace 30W fluorescent bulbs with 25W and turn off lights where no work is being done.
6. Utilize the power saving abilities on technology equipment. Many items don’t need to be running at full power 24/7, so have your staff look for opportunities (i.e., monitors can be allowed to turn off completely during non-use, new servers and power supplies will draw only the electricity they need).

The bank LEED assessment also provided numerous changes we could perform with medium to high cost investment. The payback on those investments should be measured broadly across your company’s triple bottom line.

Learn more about LEED, and about BREEAM in the EU.

Here are some of the other efficiency ideas that came out of our LEED assessment and other Green-Storming sessions:

 

Recommendation	Implementation Expectations*
	Cost	Time	LEED	ROI
			Value
Develop a policy to purchase servers with >90% efficiency power Supplies.	Low	Short	NA**	High
Replace AC power with DC to reduce heat loads.	High	Long	NA**	High
Install blanking panels.	Low	Short	NA**	Med
Install new flush valves and flow restrictors in restrooms at EDC1.	Low	Short	2	Med
Ensure all water fixtures meet the EPAct 1992 efficiency Standards.	Low	Med	0.5	Med
Install occupancy sensors on raised floor lighting.	Med	Med	1	Med
Implement vibration analysis on fans, pumps, etc. to improve efficiency.	Med	Med	NA**	Med
Use highly efficient and/or non-potable water for landscape irrigation.	Med	Med	1	Med
Mitigate storm water run-off, such as rainwater storage.	Med	Med	1	Med
Replace existing fans on computer room ACs and RTUs with premium efficiency (>92%) motors.	High	Long	NA**	Med
Install new Central Plant and chilled water distribution system.	High	Long	NA**	Med
Implement on-site, alternative energy sources, such as wind, solar, and fuel cells.	High	Long	3	Med
Provide preferred parking for carpool at sites.	Low	Short	1	Low
Purchase electricity from renewable generation sources.	Low	Med	1	Low
Shield all outdoor luminaries over 50W so no light is emitted to the night sky.	Low	Med	1	Low
Implement and document technician training and monitoring practices of O&M procedures.	Low	Short	2	Low
Add meters for humidifier and irrigation water supplies.	Low	Short	1	Low
Provide adequate outside air and monitor CO2.	Med	Med	2	Low
Add bicycle storage and changing/locker rooms.	Med	Med	1	Low
Restore nearby company owned native land.	Med	Long	0.5	Low
Replace existing transformers with Energy Star Transformers.	High	Long	NA**	Low
Confirm/improve Solar Reflectance Index of roofing material; use high emissivity and high-reflectance roofing material at roofmeeting ASTM 408 for 75% of roof.	High	Long	1	Low
Provide shade for 30% of parking lot and walkways to reduce heat island effect.	High	Long	0.5	Low

*Cost range is defined as such: Low is <$50k, Med is $50k to $200k, and High is >$200k
*Time range is defined as such: Short is <60 days, Med is 60 to 180 days, and High is >180 days
*ROI range is defined as such: Low is <1.0, Med is 1.0 to 3.0, and High is >3.0
**Recommendations all aim at improving efficiency with potential LEED point gains between 2 to 7.
 —

Next time: “Encouraging Invention and Innovation Globally.”

Green IT Innovation – Data Center Efficiencies to reduce Carbon Footprint – Part 1.

Part 1 of 3 – Consolidating Data Centers and Servers.

When most IT managers think of DC efficiency, they jump right to the topics making the most headlines, “Strategic hot and cold server rows, high efficiency power supplies, LEED certification analysis and building retrofits, etc.” Yes, I’ll cover those things in some detail in Part 3, but one of the most effective and least costly ways to cut your total electricity usage (and therefore your company’s carbon footprint) is to reduce the overall number of servers you’re using and especially strive to phase out older model servers.

In a report published by David Greenhill, Distinguished Engineer at Sun Microsystems, (Space Watts and Power), he indicates that older model servers can consume 45 times the energy of newer models. In his case study the 1997 server used 13,456 watts / hr. compared to only 300 watts / hr. for the 2005 server. But even 300 watts/hr. (0.3 Kw/hr.) add up quickly when you realize you’re typically running it 7/24/365. In fact, each 300 w server uses about 2600 Kw per year. Then you factor in the hundreds or even thousands of servers most large organizations are using and the air conditioning, UPS, lighting, etc. to house them, and you can start to see why less is definitely more.

There are plenty of other reasons for consolidating your servers too including (reduced software licensing fees, reduced maintenance costs, reduced leased space and possibly even reductions in total staff to watch over the servers.) So, the next step is figuring out How to Best Consolidate your servers.

Server Consolidation 101

• Identify and work with your key IT professionals to build a project team and plan. Select from each of the following critical areas: IT Infrastructure, Servers and Hardware, Project Management, Business Solution Managers (BSMs – point people who represent the customer), Software development, Database engineering and support, IT Architects, QA, mainframe, etc.
• Have this team build a picture of the current state: How many total servers, where are they located (country/state or province/city/building/floor), what customer groups are using which servers, what applications are on each servers, how much disk space is being used, what % of CPU utilization is being used (average and peak), etc.
• Have your IT Architects, Database, Software, BSMs and mainframe team members research to figure out which applications and databases can share the same server (not all applications are sophisticated enough to share CPU, etc.) From the list of those that cannot, work through the BSMs with your customers to see if any applications can be phased out or upgraded to enable a shared server environment.
• From the list of those applications that can live on a shared server, work with the BSMs and Project Management to initiate migration projects. (Each application and database will have its own Service Level Agreement – SLA that defines its maintenance window, etc.) and you will need to carefully coordinate with all business customers to ensure “business critical” apps are not negatively impacted during the migration.
• From your list of server utilization statistics, have your team determine the best candidate boxes to be the new primary host and backup (fail over) servers, and which boxes will be phased out (end of life’d – EOL, and hopefully recycled back to the manufacturer. See Cradle-to-Cradle. Or reused through donatation to charity organizations.)
• Decide what two physical locations will become your primary and backup data centers (more on chosing ideal DC locations in Part 2 of this series.)
• Then communicate your project plan, goals and projected benefits to all stakeholders and start implementing!
• Don’t forget to include your QA team in each step along the way and especially when verifying sytems and applications post-migration.

At the large financial institution where I participated in this type of migration, we reduced nearly 50% of our servers, saved $tens of millions off our budget and dramatically reduced our electricity and carbon footprint.

If you’re interested in learning more, please be sure to read Part 2 and 3 of this series and referencing some of the other valuable resources out there on this topic: The Green Grid, WEEE, and RoHS. Contact the author at: ToddRawlings.com.