Embracing a new “view” for data center efficiency

Recent reports have indicated that it is becoming extremely critical to optimize the management of finite data center resources, specifically power, cooling, and space. Can you explain what is driving these concerns and why enterprises should become more involved in this discussion? 

Power, cooling, and space costs are escalating year after year. It is also very difficult to scale up these elements once a facility is fully operational. Power is also a supply constrained as is space in many situations. The inability to manage these elements well introduces a higher degree of risk into the overall infrastructure and impacts the agility and sustainability as well.  The good news is that infrastructure solutions are available to help enterprises that are ready to have a conversation about adopting intelligent and sustainable technologies into their operations. These conversations should include both IT and facilities teams, as each group is a key stakeholder in efficient data center operations.

Traditionally, data center owners and managers who tend to be in the IT organization have accepted higher power and cooling costs for the sake of maximizing network uptime and overall systems and applications performance. They were chartered only with ensuring network reliability and application availability, not energy efficiency or sustainability. Energy consumption has typically only been a concern on the facilities side of the business, creating a disconnect in opportunities for maximizing efficiency. In the absence of good functional, technical and organizational alignment between IT and facilities, data centers typically are over-provisioning in terms of power, cooling and space. As a typical example, data centers tend to be over-cooled or the fill rates of the cabinets are very low, both of which inflate the cost of operations and lower efficiencies.

Additionally, enterprises need to be aware of the increasing attention that governmental agencies such as the European Union and the U.S. Environmental Protection Agency are placing on power consumption. With new governmental regulations, the emergence of carbon tax credits, and attention to corporate responsibility, effective energy management is becoming a critical performance indicator for all companies.

How do enterprises begin to address these challenges?

The first step to addressing resource challenges is to understand how the physical infrastructure can help reduce risks and costs in a way that will help a company’s bottom line, while also supporting sustainability and efficiency.

The physical infrastructure encompasses the capabilities to connect, manage and automate the different types of systems – power, control, compute, communications and security – that exist in the data center and building environment. Until recently, the physical infrastructure has not been a top of mind issue for enterprises. However, as power, cooling and space challenges as well high speed data transport (10/40/100 GB/s) and the overall performance of next generation systems technologies (Cisco Nexus, UCS, VBlock, IBM Blade Servers, iDataplex, etc.) becomes more critical, the relevance of the physical infrastructure is dramatically increasing. Today, the  physical infrastructure is the foundation for the entire network. Tomorrow, as virtualization and cloud computing become more prevalent, an intelligent and unified physical infrastructure will become a key requirement for data centers. 

Further, power, cooling, and space- related challenges are complex, cross-functional issues that can only be addressed through an integrated approach, accomplished by aligning the logical architecture with the physical infrastructure.  Technology leaders in the IT industry are introducing a range of new technologies, architectures and solutions (e.g., Cisco Nexus, UCS, VBlock and IBM Blade Center, EMC Converged Storage, etc.) to help to improve the efficiency and effectiveness of the data center environment by lowering the total cost of ownership and enhancing agility and flexibility. However, this imposes fairly stringent demands on the physical infrastructure in terms of how to effectively manage power, cooling, space, high speed data transport, manageability and scalability. The performance of the physical infrastructure is intricately tied to the performance of the next generation logical systems infrastructure– e.g., SFP+ cables in a top of row architecture to ensure low latency and power consumption.

Given the high dependence of the logical infrastructure on the physical infrastructure, it is very important that we take a holistic approach to the infrastructure. If the physical infrastructure is designed, deployed and operated aligned and integrated with the logical systems, then it will enable the logical infrastructure to become considerably more efficient and effective. This results in better application performance and improved efficiency of the data center and the entire IT organization. Software-based automation of the physical infrastructure that is integrated with the logical systems management software is a very important requirement when we look at the automation of data center operations.

Enterprises are increasingly turning to automation, along with virtualization to become more efficient. Is this an approach you would recommend to gain greater efficiencies?

The introduction of automation has done a great deal to addresses efficiency challenges, but in a very limited way. For instance, automation has proved successful in providing real-time information around such items as connectivity, which defines the relationships between connected network devices. This has enabled functions as asset tracking, automated documentation and troubleshooting, which in turn, helps streamline infrastructure operations processes and assists with capacity planning.

However, today’s automation systems must expand functionality to include energy and environmental capabilities. In addition to connectivity monitoring, systems must enable enterprises to understand what is happening from an energy consumption perspective. They must help to monitor and measure heat generation or humidity within the data centers, so that data center owners and managers have a complete end-to-end solution.

By integrating new data center infrastructure management (DCIM) capabilities  alongside traditional intelligent physical layer management (IPLM) capabilities, enterprises can monitor and catalog complete network performance and energy consumption within data centers and other IT environments, for organizations to better understand their own carbon footprint, and, put a strategy in place to reduce it.

This end-to-end monitoring of the physical infrastructure also provides the visibility, which is often lacking, to effectively execute virtualization initiatives.  In the past customers have been doing a lot of consolidation, playing with form factors of devices and merging multiple facilities into one. Now, many are adopting a virtualization approach. Yet, industry sources have estimated that only 20% of today’s virtualization is in a true production environment, while 80% is still in a development environment – meaning that most are not yet ready to risk the complete jump to virtualization. The biggest reason cited for the hesitation is lack of visibility of a management solution that can effectively monitor from an end-to-end perspective across the physical and the logical infrastructure.  This lack of visibility from a physical infrastructure perspective makes resource provisioning, capacity planning, or remote management difficult to do in an effective, energy optimized, sustainable manner.

For instance, if you have a virtual device, how do you know which physical server it is connected to and which rack that server sits in? What are the power and cooling characteristics of the rack or cabinet in which it resides? If you take virtualized servers into a full production environment and start moving them around, you begin to increase the risk of failure and downtime. Without clear visibility of these devices and the environmental conditions around them, most enterprises have hesitations about virtualization.  As we move to cloud computing, the ability to perform remote management and capacity planning with greater visibility into the entire infrastructure becomes even more critical, as well as the ability to standardize the management of the infrastructure and apply policies.

Intelligent appliances and software can help to address all of these challenges holistically, bringing them all together to provide information that can help improve operational efficiency.

How is Panduit approaching customer need for monitoring and measurement of data center performance, utilization, and energy consumption?

Panduit’s approach is to provide complete visibility of the data center infrastructure.  By monitoring real-time data, as well as the environmental factors that may impact that data, you not only have powerful information on the performance of the network, you have an appropriate level of context about that information to make more meaningful operational decisions.

Panduit’s Physical Infrastructure Management™ (PIM™) Software Platform works with PanView iQ™ (PViQ™) System Hardware to visually monitor network connectivity in real-time, offering such benefits as asset tracking, guided change procedures, and automated documentation. And the latest version of our PIM™ Software Platform is expanding these tools to enable complete data center infrastructure management, including management and monitoring of power usage, cooling temperature and humidity levels within data center racks and cabinets.  By leveraging the power of real-time information in the context of connectivity along with power, cooling and space, the PIM™ Software Platform provides meaningful information that data enterprises can use to operate more efficiently.

Let’s use an example of provisioning a server. The challenges around automating the logical provisioning of the server (obtaining the logical hardware loading the software, loading the patches, etc.) have been addressed quite satisfactorily today. The key challenge that remains unresolved is the optimization from an energy and environmental standpoint. Is it in a cabinet that has acceptable temperature characteristics? Does it have the right level of power? Is there appropriate cooling there? By understanding the context of provisioned server’s environmental factors (such as the power and thermal capacity), the server can be deployed where efficiency is optimal.  This leads to more efficient data center operations processes that are based on real-time information and are ‘energy and sustainability’ aware.

Optimization of the physical infrastructure of the data center in terms of the performance at the pod, cabinet and rack level can bring significant benefits to the power and cooling systems infrastructure, potentially lowering operational and capital expenditures. For example, if an enterprise can identify ways to improve cooling efficiency through passive means enabled by software-based automation, then it may be able to operate with fewer CRAC units, or operate at a higher thermal set point, increasing the overall efficiency of the data center operations and lowering the operating expenses.

Panduit advocates a comprehensive approach to data center operations management. Software-based automation of the physical infrastructure that is integrated with the logical systems management software will play a critical role in the optimization of data center operations as we embark on virtualization and cloud computing initiatives.