This article is a guest post from MSI-ECS, distributor of APC products.
Unveiling infrastructure secrets behind the cloud
How power and cooling reliability issues can be avoided in cloud data centers?
Thinking about the cloud? If so, be ready to recognize how cloud environments can impact data center power and cooling infrastructure. Virtualization is the technology engine behind cloud computing. Although virtualization is well understood from an IT equipment perspective, the effects of virtualization / cloud computing on the data center physical infrastructure are less understood. Some clear challenges exist that can impact systems uptime. The good news is that these challenges can be addressed by Schneider Electric™ solutions that are available today.
Challenge #1: Increased power densities
Areas of high density pose heat removal challenges. In cloud environments, virtualized servers are grouped in ways that can create “hot spots”. As a physical host is loaded up with more and more virtual machines, its CPU utilization increases, which means that more heat is produced. If the existing heat removal system is inadequate, downtime becomes a bigger threat. One solution is to spread the high density servers throughout the data center floor. However, increased floor space consumption, higher cabling costs, reduced electrical efficiency related to uncontained air paths, and the perception that half-filled racks waste rack space all could pose problems. The preferred solution is to move higher density equipment to a specific rack or row, separate from lower density equipment. This high density “pod” with dedicated heat removal (row-based cooling) and containment enables higher efficiency (shorter air paths), more predictability, and maximum density per rack. Pods and row based cooling solutions are available from Schneider Electric.
Challenge #2: Worse PUE
Virtualization can result in a significant decrease in server energy consumption. However if the power and cooling infrastructure is left in its pre-virtualization state, then power usage effectiveness (PUE) worsens. Since power is consumed by the power and cooling systems regardless of what the IT load is, when the IT load goes down (as it often does in the initial stages of virtualization), the power and cooling systems have, in effect, become oversized and less efficient. Scalable power and cooling solutions, available from Schneider Electric, can significantly reduce these inefficiencies by “rightsizing” to the new load.
Challenge #3: Increased risk of downtime
Loads within cloud data centers can vary in both time and location. Compared to a non-virtualized data center, a highly virtualized cloud-based data center will tend to experience larger load swings. Unless the cloud data center is incredibly well-planned and managed, load swings could be large enough to cause capacity-related reliability issues. This issue becomes too complex for a human to manage. Predictive management tools are a safe way to maintain an efficient, lean data center in a cloud environment. Tools such as the Schneider Electric StruxureWare™ for Data Centers suite work in real time with virtual machine managers to assure that power and cooling capacity can efficiently accommodate the dynamic loads.
Challenge #4: Unnecessary levels of redundancy
Cloud data centers are operated with a high level of IT fault-tolerance and this may reduce the need for power and cooling infrastructure redundancy. Matching of physical infrastructure redundancy to the fault-tolerant nature of a virtualized cloud environment is a form of rightsizing to further reduce energy consumption, capital costs, and fixed losses all while improving the data center PUE. Understanding the challenges that cloud computing can impose on data center physical infrastructure is an important first step in achieving the goal of a reliable, efficient and flexible data center.
For more information about high density pods, scalable power and cooling, and predictive management tools, go to www.apc.com. APC is distributed in the Philippines by MSI-ECS. For more product information, please contact email@example.com or +63-2688-3181.