Cloud Gets Complicated
The evolution from SaaS to true hybrid environments, in which cloud services are used to bring greater agility to legacy facilities, continues to advance as more organizations move to a bimodal architecture. The latest server utilization research, conducted by Stanford’s Jonathan Koomey and Anthesis Group’s Jon Taylor, found that enterprise data center servers still only deliver, on average, between five and 15 percent of their maximum computing output over the course of a year.
In addition, 30 percent of physical servers have not delivered computing services in six months or more. The push to identify and remove comatose servers will continue to build momentum and is an essential step in managing energy consumption. The potential for unused data center capacity to become part of a shared-service, distributed cloud computing model will also be explored, enabling enterprise data centers to sell their excess capacity on the open market.
Architecture Trumps Technology
While data center technology plays an important role in ensuring efficiency and availability, data center operators are focusing less on technology and more on the architectures in which those technologies are deployed.
“We are seeing more customers who in the past would have defaulted to a traditional Tier 3 or Tier 4 power architecture coming to us and asking for help in defining the right architecture for their environment,” said Peter Panfil, vice president of global power for Emerson Network Power.
Data Centers Find a Common Language
The Internet of Things (IoT) will not only impact future data center architectures by increasing the volume of data that must be processed, it will also change data center management—and the latter sooner than the former. Today’s data centers include thousands of devices that speak a host of languages, including IPMI, SNMP, and Mod Bus. This creates gaps between systems that limit efforts to manage holistically. That limit will cease to exist as Redfish, an open systems specification for data center and systems management developed by Emerson Network Power, Intel, Dell and HP, gains traction. Redfish will create interconnectivity across data center systems, enabling new levels of visibility, control and automation. Its adoption will also help establish best practices for effective use of IoT in other applications.
Social Responsibility Makes its Presence Felt
With organizations like the National Resources Defense Council raising awareness of data center energy use, some businesses are shifting their focus from efficiency to sustainability and viewing their data centers through a social responsibility lens. Data center operations—including carbon footprint, alternative energy use and equipment disposal—are now being included in corporate responsibility statements, creating greater pressure to make advances in these areas. The impact of this trend will not be limited to on-premise technology decisions. To be meaningful, reporting must include the full data center ecosystem, including colocation and cloud providers. As this practice grows, sustainability will rise to the level of availability and security as must-have attributes of a high performing data center.
Neighborhood Data Center Moves In
The growth in digital content consumption and data collection is challenging the centralized data center model. While large data centers will continue to provide the majority of computing, they will increasingly be supported by edge facilities, or neighborhood data centers, that provide low-latency content and applications to users or data processing and logic for IoT networks. As these micro data centers, operating as satellites to a central facility, proliferate on corporate campuses and in high-density residential areas, their success will depend on the use of standardized, intelligent systems that can be remotely managed.