Go ahead with that server consolidation, but make sure infrastructure measures up
IT management is under enormous pressure to maintain or increase computing and data center performance while managing costs. Server consolidation and virtualization represent one of the most promising means for accomplishing that goal. It also provides resellers with additional opportunities to offer assessment services, power, cooling and monitoring products as well as increased margins.
Server consolidation and virtualization create opportunities to reduce IT costs, enhance computing performance and improve data center space utilization. But they also introduce power and cooling challenges that, unless addressed, will limit end-user’s ability to realize the promises of these initiatives.
Is the customer’s IT infrastructure ready? End-users can get more from their virtualization initiative by identifying the gaps in their IT infrastructure and employing emerging technologies that allow smaller data centers and server rooms to achieve the flexibility and availability common within the largest facilities. Here are eight steps you can take to ensure your infrastructure is ready for a virtualized environment.
- Perform power and cooling calculations: The data center power and cooling profile will change with consolidation and virtualization. Determine power and cooling availability, capacity and redundancy for the consolidated data center to identify risks and vulnerabilities that could adversely affect the outcome of your project. Resellers can deliver a stronger ROI by removing power and cooling constraints that limit end-users ability to add new higher density servers in less space and reduce energy costs through more efficient power and cooling systems.
Calculate the power required by the data center and within each individual rack to ensure adequate UPS capacity. Look for potential points of failure, where the loss of one power component can threaten the larger system. Keeping power line drawings for the room up-to-date can allow you to identify where to employ redundancy.
For cooling analysis, a Computational Fluid Dynamics (CFD) assessment can show exactly how airflow will occur in the consolidated environment and where hot spots and other cooling challenges will exist. Your room and rack power loads can help determine levels of cooling capacity. Redundancy can be built into the cooling system to eliminate potential points of failure. - Provide physical security: In the more critical consolidated environment, equipment needs to be protected from unauthorized access--whether your data center is a single rack or multiple racks.
Utilize racks that have key or card swipe locks so you can authorize access only by qualified individuals. Within the racks, utilize smart PDUs, which let you monitor power at the receptacle level and enable receptacles to be activated or turned off. This will help prevent unauthorized equipment adds and changes that can lead to circuit overloads, as well as simplifying in-rack power distribution and reducing cabling. - Ensure you have adequate UPS capacity: Industry analyst and author Barb Goldworm summed up the reality that should shape your critical power strategy: “With virtualization, since each server runs multiple virtual server workloads, the server becomes a single point of failure. This escalates the issue of availability for that physical server. If you put all your virtual eggs in one basket, then you need to take very good care of that basket.”
UPSs installed years ago may not have the capacity or reliability required for upgraded servers or blade servers common to consolidation / virtualization projects. Ensure that UPS sizing and backup are adequate for today’s needs and future growth.
Make sure you calculate your UPS size based on the full load of protected equipment--not “nominal loads,” which are estimates of average loads and could result in under sizing your UPS capacity.
Power protection, like computing, can benefit from consolidation of multiple smaller systems into fewer, more powerful systems. When total room load exceeds 15kW, consider replacing rack-based UPS systems with room level systems. - Increase the reliability of your UPSs: You may desire to use the same line interactive UPSs you’ve always used for racks. However, the consolidated environment increases criticality, and may be better served by online UPSs with fault tolerant design. An online rack UPS can be twice as reliable as more commonly used line interactive UPS, as measured in Mean Time Between Failure (MTBF). That’s because it has an internal bypass that allows power to continue to protect equipment in the event of a UPS component failure. Room-level UPSs have even higher reliability due to their bypass technology and longer lasting batteries.
- Plan for power redundancy: The critical environment of consolidation and virtualization will likely mean you want to add redundancy to your power system. The servers you use will likely be dual corded, which means they are designed for power redundancy. The highest level of redundancy is providing separate UPS, PDU and circuit for each cord. Exactly how much power capacity and redundancy you need depends on your organization’s sensitivity to downtime and your budget.
- Increase the criticality of your cooling systems: If the heat from a rack is not effectively removed, the performance, availability and lifespan of the equipment in the rack will be reduced significantly. As you adopt the latest server technologies in your data center, you may be exposed to higher failure rates, especially in the top third of the racks where heat builds up.
The right cooling systems can eliminate the constraints to server deployment caused by higher temperatures, allowing you to fully populate your racks. If you are using building air conditioning for your IT spaces, consolidation will likely create a need for dedicated cooling and even high-density cooling. Compared to comfort cooling, dedicated systems provide year round cooling, move more air than comfort systems, eliminate large swings in temperature and humidity and provide higher levels of air filtration. For spaces with single racks, employ an integrated cooling enclosure that can support power loads of up to 1.6kW and an enclosure system able to sustain high-density power loads of up to 14.4kW per rack.
For high-density situations where room-level precision cooling is already in place, consider supplemental high-density cooling. These systems bring cooling closer to the rack and therefore provide an effective and efficient solution for high density issues. - Gain real-time visibility and control: With server consolidation and virtualization, the need to know what is happening within equipment racks at any given time becomes vital. Ensure your UPS systems are network enabled, using monitoring technology will provide alerts for adverse system or environmental conditions, and allow you to conduct graceful shutdowns of IT equipment in the event of extended power outages.
- Upgrade service and maintenance: The addition of higher criticality IT equipment is an ideal time to revisit your service and maintenance processes and contracts.
Extending the useful service life of the power and cooling equipment through proper maintenance, predictive monitoring, and keeping the equipment up-to-date increases equipment life spans and maximizes performance.
Service level agreements should be reviewed at this time and modified with higher criticality in mind. After adding new power and cooling equipment, resellers should be sure to offer extended warranties and maintenance contracts to ensure performance and a full lifespan. End-users should also insure their investment by using factory certified service technicians for maintenance.
Be prepared for virtualization. Starting off on the right foot is the best way to make sure that your infrastructure is able to adapt to virtualized environment. By making some upfront calculations and following the right steps, your business will be ready for the change.
Tom Karabinos is the Director of Partner Channels for Emerson Network Power/ Liebert Corporation, responsible for all aspects of the Liebert Network Solutions Partner (LSNP) program, which involves Solution Providers, VARs, Systems Integrators and major distributors.
Tom has more than 25 years of experience in the critical support systems market, which includes power backup (UPS), power distribution, precision cooling, transient voltage surge suppression (TVSS), racks and enclosures, and system monitoring of data centers, computer rooms and network systems. Contact Tom at 614-841-8166.
