TY - JOUR
T1 - Processor-embedded distributed smart disks for I/O-intensive workloads
T2 - Architectures, performance models and evaluation
AU - Chiu, Steve C.
AU - Liao, Wei Keng
AU - Choudhary, Alok N.
AU - Kandemir, Mahmut T.
N1 - Funding Information:
Alok N. Choudhary is a professor in the Electrical and Computer Engineering Department at Northwestern University. He also holds an appointment with the Kellogg Graduate School of Management in the Marketing and Technology Innovation Departments. Alok Choudhary received his Ph.D. from the University of Illinois, Urbana-Champaign, in Electrical and Computer Engineering, in 1989, an M.S. from University of Massachusetts, Amherst, in 1986, and B.E. (Hons.) from Birla Institute of Technology and Science, Pilani, India in 1982. He received the National Science Foundation's Young Investigator Award in 1993, an IEEE Engineering Foundation award, on IBM Faculty Development award, and an Intel Research Council award. Dr. Choudhary serves on the editorial boards of IEEE Transactions on Parallel and Distributed Systems, Journal of Parallel and Distributed Computing, and International Journal of High Performance Computing and Networking.
PY - 2005/4
Y1 - 2005/4
N2 - Processor-embedded disks, or smart disks, with their network interface controller, can in effect be viewed as processing elements with on-disk memory and secondary storage. The data sizes and access patterns of today's large I/O-intensive workloads require architectures whose processing power scales with increased storage capacity. To address this concern, we propose and evaluate disk-based distributed smart storage architectures. Based on analytically derived performance models, our evaluation with representative workloads show that offloading processing and performing point-to-point data communication improve performance over centralized architectures. Our results also demonstrate that distributed smart disk systems exhibit desirable scalability and can efficiently handle I/O-intensive workloads, such as commercial decision support database (TPC-H) queries, association rules mining, data clustering, and two-dimensional fast Fourier transform, among others.
AB - Processor-embedded disks, or smart disks, with their network interface controller, can in effect be viewed as processing elements with on-disk memory and secondary storage. The data sizes and access patterns of today's large I/O-intensive workloads require architectures whose processing power scales with increased storage capacity. To address this concern, we propose and evaluate disk-based distributed smart storage architectures. Based on analytically derived performance models, our evaluation with representative workloads show that offloading processing and performing point-to-point data communication improve performance over centralized architectures. Our results also demonstrate that distributed smart disk systems exhibit desirable scalability and can efficiently handle I/O-intensive workloads, such as commercial decision support database (TPC-H) queries, association rules mining, data clustering, and two-dimensional fast Fourier transform, among others.
UR - http://www.scopus.com/inward/record.url?scp=14944352193&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=14944352193&partnerID=8YFLogxK
U2 - 10.1016/j.jpdc.2004.01.006
DO - 10.1016/j.jpdc.2004.01.006
M3 - Comment/debate
AN - SCOPUS:14944352193
SN - 0743-7315
VL - 65
SP - 532
EP - 551
JO - Journal of Parallel and Distributed Computing
JF - Journal of Parallel and Distributed Computing
IS - 4
ER -