Computer Systems Research  (CSR)

SYNOPSIS

Advances in software and hardware technologies are expanding the frontiers of distributed computing along multiple dimensions. Progress in Internet technologies is revolutionizing the use and scale of distributed systems, ushering in a variety of applications, such as Web search, e-commerce, social networking, distributed virtual environments, and utility computing. Advances in handheld and embedded device technology, coupled with progress in wireless communications and mobile computing, have led to the emergence of pervasive and ubiquitous computing where technology is increasingly invisible and where access to information and services is provided unobtrusively anytime, anywhere. Recent advances in multi-core architectures is enabling new levels of parallelism, never seen before in mainstream computing, with potential for significant performance gains without exacerbating the problems of power dissipation and design complexity. The Computer Systems Research (CSR) program supports the exploration of the new frontiers of computer systems and software, focusing on systems research that explores novel ideas and expands the limits of existing paradigms, with potential for significant advances in scientific or technical understanding of future computing systems and applications.

The need to operate in heterogeneous, unpredictable and challenging environments requires ground-breaking approaches and methodologies to advance our understanding of how computation is performed and how resources are managed, at varying levels of granularity and scale. The proliferation of Internet-scale applications and services poses new challenges and require radical thinking of how future file and storage systems are designed and managed. The difficulty of these challenges grows with the number of users and the intensity of the data. This is further compounded by the need for energy-efficient and self-managing storage capabilities, support for pervasive access to personal storage, and support for caching, replication and consistency. Frameworks, approaches and methodologies to address these challenges must show potential to improve system’s characteristics, such as manageability, configurability, operational sustainability, usability and performance, while reducing vulnerabilities.

As mobile device technology continues to evolve, pervasiveness and ubiquity are increasingly becoming essential requirements of future distributed systems. The dynamic and heterogeneous nature of ubiquitous and pervasive computing environments, coupled with the interaction between human and devices, give rise to unique fundamental and socio-technical challenges. At the core of these challenges is the concept of context, its representation and the underlying principles that underpin how human behavior, activity and interaction with the environment are captured at the appropriate levels of detail. Advances in context-aware, pervasive and ubiquitous computing require new programming models, abstractions and languages. Methodologies and tools are also needed to monitor, evaluate and predict the performance of ubiquitous systems and assess users’ experience. Collaborations with researchers in artificial intelligence and the social sciences that provide new perspectives on how human and context-aware ubiquitous computing are encouraged.

Fully leveraging the opportunities and unprecedented levels of parallelism offered by multi-core architectures poses new challenges which bring into question traditional frameworks, approaches and methodologies for system and software design in large-scale, high performance environments. Addressing these challenges requires sound parallel execution and memory models, innovative system-level approaches to automatic parallelization of sequential programs, novel compiler techniques and dynamic run-time execution to expose and exploit inherent parallelism and optimize code generation, and new design approaches for high performance I/O systems. Understanding parallel systems and applications also requires innovative methodologies and tools for quantitative and qualitative characterization, evaluation, monitoring and prediction of system behavior at different levels, including the implications of workloads in system design in large-scale, high performance environments.

CSR seeks advances that are specific to an application domain or a particular hardware platform as well as generic across domains and/or platforms. Also sought are proposals focused on advancing the state-of-the art in systems and software research for compute-intensive applications and hardware. Proposal focused on data-intensive applications and hardware should be submitted to the Data-intensive Computing cross-cutting program <insert link>. Investigators interested in the CSR program may also wish to consider the Software and Hardware Foundations program <insert link>, which supports foundational software and hardware research essential to enhance the capability of computing systems. CSR PIs should describe credible plans for demonstrating the utility and potential impact of their proposed work.

For more information on the types of projects supported by the CSR program, please visit our web site at http://www.nsf.gov/cise/cns/csr_pgm.jsp. 

Computer Systems Research (CSR) Staff

Funding Opportunities for the Computer Systems Research (CSR) Program:

Computer and Network Systems (CNS): Core Programs.  NSF 08-576

THIS PROGRAM IS PART OF

Computer and Network Systems (CNS): Core Programs