As the scale of data has grown from the order of megabytes to the order of petabytes, energy and power have become a key consideration in managing (i.e., analyzing, organizing, and storing) such large datasets due to the increasing cost of powering the required data processing resources and cooling equipment. It is expected that the demand for data processing resources, both in the form of servers and mobile computing devices, will continue to increase exponentially. The goal of this two-day workshop is to act as a planning meeting for the development and deployment of energy-efficient data management methods and data-intensive applications which are important from both an economic and a sustainability perspective. The workshop includes invited talks, panels and breakout sessions.
The broader impact of the workshop is expected to be in terms of guiding future research which will provide a more sustainable way of managing society's increasing and rapidly growing reliance on large data management systems.
More information on the workshop can be found at www.energy-efficient-data-management.org/workshop.
The total energy used by the Information Technology (IT) industry is currently a few percent of the total energy usage in the United States, roughly comparable to the energy used by the airline industry; it is in fact expected to grow even further in the near future. This is explained by two converging trends. The first trend is that the total energy used by computing technology, and the energy density within the technology, has been increasing exponentially for decades. The second trend is that the number of computing devices, particularly mobile computing devices, has been also increasing exponentially for a similar time frame. Further, it has been estimated that energy-related costs account for about a third of the total cost-of-ownership of data processing and computing systems. Thus, economic reasons as well as the adoption of environmentally friendly practices (due to the increasing awareness of environmental problems) has led to this rapid emergence of power and energy as a first-class citizen in IT. This clearly suggests that we need to re-examine the existing data management and processing approaches at all levels. The first NSF Workshop on Sustainable Energy-efficient Data Management (SEEDM 2011) was held on May 1-3, 2011 in Arlington, VA with the goal of to act as a planning meeting for the development and deployment of sustainable energy-efficient data management and data-intensive applications. The 38 participants of the workshop (who were selected based on their submitted position papers) included well-established senior researchers, new emerging researchers, and advanced developers from both academia and industry involved in all aspects of data management in diverse environments. There were also representatives from government labs and selected government agencies, as well as two participants from Europe. The workshop included a balance of presentations, panels, and group discussions. The invited talks were carefully selected to set up the stage for the ensuing break-out group discussions. These included talks on academic and industrial research, as well as reports from other related NSF-funded workshops. The workshop participants have identified that data management and information technology can contribute to energy sustainability in two ways. Directly, through Energy-Efficient Data Management, i.e., by processing Big Data with less energy. Indirectly, through Data-Centric Energy Management, i.e., by processing data in a timely manner to support sustainable energy policies that 'use less (right provisioning for expected and reserve vs. peak), use what is needed (power proportionality), use better (integrate renewable energies).' During the discussion three things became clear: Power and energy should be a first-order resource and system design principle across all levels, instead of an afterthought. Their importance should also be made apparent through updated curricula, starting at the undergraduate level. Energy-efficient data management can be achieved with enabling hardware and cross-layer optimizations. Thus, stronger interaction among the data management, computer architecture, and data storage communities is needed in order for significant energy-efficiency in data management to be achieved. Continuous Data analytics (data stream processing and on the-fly data mining) and data visualization techniques are fundamental in meeting the goal of a sustainable energy grid. A starting point would be the design of smart (information-aware) buildings and communities. Finally, a shared belief among all the participants is that NSF can play an important role in advancing energy-efficient computing, by facilitating energy-related infrastructure deployment and experimentations (e.g., through a community repository for energy data, along the lines of other disciplines such as astronomy, biology, etc). The main web site of the workshop, located at www.seedm.org, includes logistics information for participants, complete list of participants, and a detailed schedule of activities. It also includes the position papers of all participants that form the workshop proceedings as well as the slides and notes from all the presentations.
