Smart cars, smart homes, smart devices… embedded intelligence has changed the way we live and redefined industry methodologies around the world. Connecting and empowering these devices puts skyrocketing stress on the IT teams that manage data centers.
Challenge 1: evolving the Internet
The Internet originated to connect people and has gone through numerous changes to keep up with the growth of the user base as well as today’s smart devices and apps that enable flexible, anywhere accessibility. People want 24x7 connections with all of their devices and data. According to a Nielsen studyd, adult smart phone users spend 65 percent more time each month using apps compared to two years prior.
This activity contributes to the skyrocketing growth of Internet traffic hitting data centers. Facebook, for example, in August of this year experienced 1 billion site visitors in a single day.
While the Internet continues to connect more people every day, the profound growth of traffic hitting data centers stems from operator-less smart devices deployed in industrial environments. Two thirds of the global economy today relies on network devices. Gartner Group estimates that 25-billion connected “Things” will be in use by 2020, and IDC’s prediction for the “Internet of Things” puts the number of installed smart devices at 30 billion in that same time frame.
Challenge 2: building out data centers
More devices mean more traffic and more demands on the data center. IT teams are used to continually striving to be more efficient, and being able to do more with less. Keeping up in the world of the Internet of Things, however, has proven that the usual methodologies are not enough. Virtualization and consolidation have become the norm, and still the demands on servers continue to rise.
The challenges of keeping up with demands start with scaling. As the infrastructure expands to include more devices, more servers are inevitably required. In the case of smart phones alone, a new server must be added for approximately every 600 phones. Besides scaling issues, IT teams face the challenges of increasingly dynamic environments. Besides high volumes of traffic, the types of requests and services translate into very diverse workloads that are difficult to predict.
Power has also become a challenge for the world’s largest data centers as well as sites where local utilities put restrictions on energy availability or quality. Data centers, in fact, consume a significant and increasing percentage of the world’s overall energy supplies. As a result, energy costs have become the fastest growing component of data center OpEx.
Smarter data centers
Energy will always be on the list of data center challenges, however it should also be included in the list of factors that can be harnessed to accomplish far more than just manage the costs of power and cooling. As the lifeblood of the data center, energy can be monitored to gain insights about activity levels, workload dynamics, potential threats to continuity, and targets for optimization and efficiency improvements.
In fact, energy management is at the heart of a more agile IT approach in the data center.
By tapping into the existing temperature and power consumption sensors built into modern data center equipment, IT and facilities teams can work smarter. Middleware solutions that gather and aggregate the real-time data points enable more intelligent oversight, informed decisions, and a range of automated functions that benefit operations.
Consider the types of high-level and fine-grained visibility and control enabled by energy management middleware:
- Real-time monitoring, with thermal and energy maps that highlight hot spots and under-utilized resources
- Trending analysis, drawing from actual history and not theoretical estimations
- Power allocation and capping
- Dynamic management policy implementation (prioritize/automate energy allocation)
- Higher-density data centers (fewer servers; more servers per rack; more racks per room)
- Reduced energy consumption (removal of under-utilized resources; power capping to avoid going over energy budgets without compromising performance)
- Ability to safely increase operating temperature (lower cooling costs) and introduce advanced cooling approaches
- Replace expensive smart power strips with middleware (software) instrumentation
Flagship data centers continue to expand this list of use cases, and energy monitoring has become a standard tool for many of the IT teams that deliver services to the Internet of Things.
Smart deployment options
Energy management solutions vary immensely, and can be confusing when considered within the larger framework of a full-feature data center infrastructure management solution. The hype around DCIM has yet to yield mature, versatile, and affordable solutions. And the complexity of DCIM deployments makes this class of solutions far too daunting for many organizations. Months of integration can be required to tie a DCIM solution into an organization’s existing system of record, information service management platform, or configuration management database.
Note, however, that energy management middleware can be introduced independently from a DCIM platform. Data center managers can choose an easy-to-deploy energy management solution today, and benefit from almost immediate ROI in the form of lower utility bills.
Besides researching stand-alone energy management tools, IT teams should reach out to their usual circle of integrators and solution suppliers. Many dashboard and console vendors are offering fully integrated energy management functionality.
The choice of deployment options reflects the growing demand for energy management functionality, and the increasing reliance on temperature and power monitoring and control as a foundation for improved data center management. The Internet of Things is here to stay, and energy management makes it possible to very effectively align data centers with the users and devices that they ultimately serve.
To learn more visit www.intel.com/dcm, and follow us on Twitter