The Experts below are selected from a list of 978 Experts worldwide ranked by ideXlab platform
S C Liew - One of the best experts on this subject based on the ideXlab platform.
-
performance of various input buffered and output buffered atm switch design principles under bursty Traffic simulation study
IEEE Transactions on Communications, 1994Co-Authors: S C LiewAbstract:This paper investigates the packet loss probabilities of several alternative input-buffered and output-buffered switch designs with finite amounts of buffer space. The effects of bursty Traffic, modeled by geometrically distributed active and idle periods, are explored. Methods for improving switch performance are classified, and their effectiveness for dealing with bursty Traffic discussed. This work indicates that bursty Traffic can degrade switch performance significantly and that it is difficult to alleviate the performance degradation by merely restricting the offered Traffic load. Unless buffers are shared, or very large buffers provided, strategies that improve throughput under Uniform Random Traffic are not very effective under bursty Traffic. For input-buffered switches, our investigation suggests that the specific contention resolution scheme we use is a more important performance factor under bursty Traffic than it is under Uniform Random Traffic. In addition, many qualitative results true for Uniform Random Traffic are not true for bursty Traffic. The work also reveals several interesting, and perhaps unexpected, results: 1) output queueing may have higher loss probabilities than input queueing under bursty Traffic; 2) speeding up the switch operation could result in worse performance than having several output ports per output address under bursty Traffic; and 3) if buffers are not shared in a fair manner, sharing buffers could make performance worse than not sharing buffers at high Traffic loads. Simulation results and intuitive explanations supporting the above observations are presented. >
John Kim - One of the best experts on this subject based on the ideXlab platform.
-
indirect adaptive routing on large scale interconnection networks
International Symposium on Computer Architecture, 2009Co-Authors: Nan Jiang, John KimAbstract:Recently proposed high-radix interconnection networks [10] require global adaptive routing to achieve optimum performance. Existing direct adaptive routing methods are slow to sense congestion remote from the source router and hence misroute many packets before such congestion is detected. This paper introduces indirect global adaptive routing (IAR) in which the adaptive routing decision uses information that is not directly available at the source router. We describe four IAR routing methods: credit round trip (CRT) [10], progressive adaptive routing (PAR), piggyback routing (PB), and reservation routing (RES). We evaluate each of these methods on the dragonfly topology under both steady-state and transient loads. Our results show that PB, PAR, and CRT all achieve good performance. PB provides the best absolute performance, with 2-7% lower latency on steady-state Uniform Random Traffic at 70% load, while PAR provides the fastest response on transient loads. We also evaluate the implementation costs of the indirect adaptive routing methods and show that PB has the lowest implementation cost requiring <1% increase in the total storage of a typical high-radix router.
-
indirect adaptive routing on large scale interconnection networks
International Symposium on Computer Architecture, 2009Co-Authors: Nan Jiang, John Kim, William J DallyAbstract:Recently proposed high-radix interconnection networks [10] require global adaptive routing to achieve optimum performance. Existing direct adaptive routing methods are slow to sense congestion remote from the source router and hence misroute many packets before such congestion is detected. This paper introduces indirect global adaptive routing (IAR) in which the adaptive routing decision uses information that is not directly available at the source router. We describe four IAR routing methods: credit round trip (CRT) [10], progressive adaptive routing (PAR), piggyback routing (PB), and reservation routing (RES). We evaluate each of these methods on the dragonfly topology under both steady-state and transient loads. Our results show that PB, PAR, and CRT all achieve good performance. PB provides the best absolute performance, with 2-7% lower latency on steady-state Uniform Random Traffic at 70% load, while PAR provides the fastest response on transient loads. We also evaluate the implementation costs of the indirect adaptive routing methods and show that PB has the lowest implementation cost requiring
Nan Jiang - One of the best experts on this subject based on the ideXlab platform.
-
indirect adaptive routing on large scale interconnection networks
International Symposium on Computer Architecture, 2009Co-Authors: Nan Jiang, John KimAbstract:Recently proposed high-radix interconnection networks [10] require global adaptive routing to achieve optimum performance. Existing direct adaptive routing methods are slow to sense congestion remote from the source router and hence misroute many packets before such congestion is detected. This paper introduces indirect global adaptive routing (IAR) in which the adaptive routing decision uses information that is not directly available at the source router. We describe four IAR routing methods: credit round trip (CRT) [10], progressive adaptive routing (PAR), piggyback routing (PB), and reservation routing (RES). We evaluate each of these methods on the dragonfly topology under both steady-state and transient loads. Our results show that PB, PAR, and CRT all achieve good performance. PB provides the best absolute performance, with 2-7% lower latency on steady-state Uniform Random Traffic at 70% load, while PAR provides the fastest response on transient loads. We also evaluate the implementation costs of the indirect adaptive routing methods and show that PB has the lowest implementation cost requiring <1% increase in the total storage of a typical high-radix router.
-
indirect adaptive routing on large scale interconnection networks
International Symposium on Computer Architecture, 2009Co-Authors: Nan Jiang, John Kim, William J DallyAbstract:Recently proposed high-radix interconnection networks [10] require global adaptive routing to achieve optimum performance. Existing direct adaptive routing methods are slow to sense congestion remote from the source router and hence misroute many packets before such congestion is detected. This paper introduces indirect global adaptive routing (IAR) in which the adaptive routing decision uses information that is not directly available at the source router. We describe four IAR routing methods: credit round trip (CRT) [10], progressive adaptive routing (PAR), piggyback routing (PB), and reservation routing (RES). We evaluate each of these methods on the dragonfly topology under both steady-state and transient loads. Our results show that PB, PAR, and CRT all achieve good performance. PB provides the best absolute performance, with 2-7% lower latency on steady-state Uniform Random Traffic at 70% load, while PAR provides the fastest response on transient loads. We also evaluate the implementation costs of the indirect adaptive routing methods and show that PB has the lowest implementation cost requiring
William J Dally - One of the best experts on this subject based on the ideXlab platform.
-
indirect adaptive routing on large scale interconnection networks
International Symposium on Computer Architecture, 2009Co-Authors: Nan Jiang, John Kim, William J DallyAbstract:Recently proposed high-radix interconnection networks [10] require global adaptive routing to achieve optimum performance. Existing direct adaptive routing methods are slow to sense congestion remote from the source router and hence misroute many packets before such congestion is detected. This paper introduces indirect global adaptive routing (IAR) in which the adaptive routing decision uses information that is not directly available at the source router. We describe four IAR routing methods: credit round trip (CRT) [10], progressive adaptive routing (PAR), piggyback routing (PB), and reservation routing (RES). We evaluate each of these methods on the dragonfly topology under both steady-state and transient loads. Our results show that PB, PAR, and CRT all achieve good performance. PB provides the best absolute performance, with 2-7% lower latency on steady-state Uniform Random Traffic at 70% load, while PAR provides the fastest response on transient loads. We also evaluate the implementation costs of the indirect adaptive routing methods and show that PB has the lowest implementation cost requiring
Yuliang Li - One of the best experts on this subject based on the ideXlab platform.
-
a scalable low latency high throughput optical interconnect architecture based on arrayed waveguide grating routers
Journal of Lightwave Technology, 2015Co-Authors: Roberto Proietti, Christopher Nitta, Yuliang LiAbstract:This paper proposes, simulates, and experimentally demonstrates an optical interconnect architecture for large-scale computing systems. The proposed architecture, Hierarchical Lightwave Optical Interconnect Network (H-LION), leverages wavelength routing in arrayed waveguide grating routers (AWGRs), and computing nodes (or servers) with embedded routers and wavelength-specific optical I/Os. Within the racks and clusters, the interconnect topology is hierarchical all-to-all exploiting passive AWGRs. For the intercluster communication, the proposed architecture exploits a flat and distributed Thin-CLOS topology based on AWGR-based optical switches. H-LION can scale beyond 100 000 nodes while guaranteeing up to 1.83×saving in number of inter-rack cables, and up to 1.5×saving in number of inter-rack switches, when compared with a legacy three-tier Fat Tree network. Network simulation results show a system-wide network throughput reaching as high as 90% of the total possible capacity in case of synthetic Traffic with Uniform Random distribution. Experiments show 97% intracluster throughput for Uniform Random Traffic, and error-free intercluster communication at 10 Gb/s.
