Cloud computing has emerged as a dominant computing platform providing billions of users world-wide with online services. The software applications powering these services, commonly referred to as scale-out workloads and which include web search, social networking and business analytics, tend to be characterized by massive working sets, high degrees of parallelism, and real-time constraints – features that set them apart from desktop, parallel and traditional commercial server applications. To support the growing popularity and continued expansion of cloud services, providers must overcome the physical space and power constraints that limit the growth of data centers. Problematically, the predominant processor micro-architecture is inherently inefficient for running these demanding scale-out workloads, which results in low compute density and poor trade-offs between performance and energy. Continuing the current trends for data production and analysis will further exacerbate these inefficiencies.
Improving the cloud’s computational resources whilst operating within physical constraints requires server efficiency to be optimized in order to ensure that server hardware meets the needs of scale-out workloads. To this end, the team of Babak Falsafi, a Professor in the School of Computer and Communication Sciences at EPFL, the director of the EcoCloud research center at EPFL (founded to innovate future energy-efficient and environmentally friendly cloud technologies) and a HiPEAC member, presented Clearing the Clouds: A Study of Emerging Workloads on Modern Hardware, which received the best paper award at ASPLOS 2012.
In this paper, the EPFL team explained how they used performance counters on modern servers to assess how well today’s predominant processor micro-architecture is aligned with the requirements of scale-out applications. What they discovered is that there is a significant mismatch between the two, stemming from inefficiencies in the instruction supply and execution logic as well as memory system organization. Their research shows that efficiently executing scale-out workloads requires optimizing the instruction-fetch path for multi-megabyte instruction working sets, reducing the core complexity, and shrinking the capacity of on-die caches to reduce area and power overheads. The authors also introduced CloudSuite, a benchmark suite of emerging scale-out workloads, that is expected to benefit the broader research community.
The insights gleaned as part of the evaluation are now driving the team to develop server processors tuned to the demands of scale-out workloads. The team has recently proposed a processor organization that unlike current industrial chip design trends does away with power-hungry cores and large on-die caches and networks to free area and power for a large number of simple cores built around a streamlined memory hierarchy. Not only do these improvements lead to greater performance and efficiency at the level of each processor chip, they also enable significant cost and power savings at the level of an entire data center.
This work was partially funded by the EuroCloud Server Project, a European Commission FP7 Computing Systems Program and is deemed as a European “flagship” project, led by major research centers and industrial partners such as ARM, IMED, Nokia and the University of Cyprus. Running from Jan-2010 until Dec-2012, EuroCloud’s multiple partners are focused on increasing by 10x the efficiency in server chip level power consumption. Dr. Max Lemke, Deputy Head of Unit for Embedded Systems and Control in the Directorate General Information Society and Media of the European Commission, referred to the project to illustrate how the main goal of research in computing systems is getting energy efficient and low-cost computing technologies into the full spectrum of devices and systems, from mobile and embedded systems to data centers and supercomputers.
“Computing is a key enabler for Europe’s competitiveness in engineering, which is a key driver for the European economy,” Dr. Lemke said in his keynote address at the recent HiPEAC 2012 Conference. “Europe has to leverage its unique expertise in embedded and mobile computing systems to innovate in energy efficient and low-cost computing technologies,” he added.
Babak joined the School of Computer and Communication Sciences at EPFL in 2008. Prior to that, he was a full Professor of Electrical & Computer Engineering and Computer Science at Carnegie Mellon where he led the Microarchitecture theme of the FCRP Center on Circuit and System Solutions, a multi-university consortium of over 50 academics investigating digital platform designs for the end of CMOS roadmap. He is the founding director of the EcoCloud research center pioneering future energy-efficient and environmentally-friendly cloud technologies at EPFL.
His research targets technology-scalable datacenters, design for dark silicon, architectural support for software and hardware robustness, and analytic and simulation tools for computer system performance evaluation. He is a recipient of an NSF CAREER award in 2000, IBM Faculty Partnership Awards in 2001, 2003 and 2004, and an Alfred P. Sloan Research Fellowship in 2004. He has been a member of ISCA Hall of Fame since 2003 and the Micro Hall of Fame since 2011 for contributions to the flagship IEEE/ACM conferences in computer architecture and microarchitecture respectively. He is a fellow of IEEE.