|PINCIROLI RICCARDO||Cycle: XXX |
Section: Computer Science and Engineering
Tutor: PERNICI BARBARA Major Research topic
:Energy efficiency in large data centers using performance evaluation techniques
Advisor: GRIBAUDO MARCOAbstract:
Nowadays, the amount of available computing resources is limitless and the energy consumption in IT is assuming an increasing importance due to the huge number of compute-intensive applications used in business and scientific fields. It has been estimated that by 2020, the U.S. data centers demand for electricity will increase to 140 billion kWh annually; it means that American enterprises will pay $13 billion annually in electricity bills, emitting almost 100 million metric tons of carbon pollution per year. While several efforts have been done in order to improve the energy efficiency, it is evident that the energy consumption must be further reduced, taking into consideration the heterogeneity of the current applications. In my work, I will propose new techniques to evaluate the energy efficiency of data centers. To this end, I will apply performance modeling approaches such as queuing networks and stochastic processes, using analytical and discrete event simulation techniques. In particular, I will propose new metrics in order to better analyze the energy efficiency of data centers. Indeed, the actual metrics do not consider the requirements of different application, an important feature which determines how resources will be used by requests. The characterization of the different applications in terms of resource requirements is the key feature that will be considered in the development of the thesis. The main objective that will be pursued is to decrease the data centers energy consumption, still considering other important aspects, such as the SLAs stipulated with customers.