Hamid Fadishei, Hossein Deldari, and Mahmoud Naghibzadeh. "Pre-execution power consumption prediction of computational multithreaded workloads." Cluster Computing vol. 17, no. pp. 1323-1333, 2014, http://dx.doi.org/10.1007/s10586-014-0401-0
Abstract: Power management in large-scale computational environments can significantly benefit from predictive models. Such models provide information about the power consumption behavior of workloads prior to running them. Power consumption depends on the characteristics of both the machine and the workload. However, combinational features such as the cache miss rate cannot be considered due to their unavailability before running the workload. Therefore, pre-execution power modeling requires both machine-independent workload characteristics and workload-independent machine characteristics. In this paper the predictive modeling problem is tackled by the proposal of a two-stage modeling framework. In the first stage, a machine learning approach is taken to predict single-threaded workload power consumption at a specific frequency. The second stage analytically scales this output to any intended thread/frequency configuration. Experimental results show that the proposed approach can yield highly accurate predictions about workload power consumption with an average error of 3.7 % on six different test platforms.
Keywords: Power-aware computing, Computational power modeling, Abstract workload modeling, Simultaneous multithreading
Hamid Saadatfar, Hamid Fadishei, and Hossein Deldari. "Predicting Job Failures in AuverGrid Based on Workload Log Analysis." New Generation Computing 30.1 (2012): 73-94, http://dx.doi.org/10.1007/s00354-012-0105-z
Abstract: Grid systems are popular today due to their ability to solve large problems in business and science. Job failures which are inherent in any computational environment are more common in grids due to their dynamic and complex nature. Furthermore, traditional methods for job failure recovery have proven costly and thus a need to shift toward proactive and predictive management strategies is necessary in such systems. In this paper, an innovative effort has been made to predict the futurity of jobs in a production grid environment. First of all, we investigated the relationship between workload characteristics and job failures by analyzing workload traces of AuverGrid which is a part of EGEE (Enabling Grids for E-science) project. After the recognition of failure patterns, the success or failure status of jobs during 6 months of AuverGrid activity was predicted with approximately 96% accuracy. The quality of services on the grid can be improved by integrating the result of this work into management services like scheduling and monitoring.
Keywords: Job Failure Prediction, Grid Workload Archive, Trace Analysis, Bayesian Networks
Hamid Fadishei, Hamid Saadatfar, Hossein Deldari, "Job failure prediction in grid environment based on workload characteristics," Proceedings of the 14th International CSI Computer Conference, CSICC 2009, pp. 329-334, 20-21 Oct. 2009, http://dx.doi.org/10.1109/CSICC.2009.5349381
Abstract: The power of grid technology in aggregating autonomous resources owned by several organizations into a single virtual system has made it popular in compute-intensive and data-intensive applications. Complex and dynamic nature of grid makes failure of users' jobs fairly probable. Furthermore, traditional methods for job failure recovery have proven costly and thus a need to shift toward proactive and predictive management strategies is necessary in such systems. In this paper, an innovative effort is made to predict the futurity of jobs submitted to a production grid environment (AuverGrid). By analyzing grid workload traces and extracting patterns describing common failure characteristics, the success or failure status of jobs during 6 months of AuverGrid activity was predicted with around 96% accuracy. The quality of services on grid can be improved by integrating the result of this work into management services like scheduling and monitoring.
Hamid Saadatfar, Hamid Fadishei, and Hossein Deldari. "The Study of the Relations Between Grid Job Failure Patterns and Workload Characteristics." IEEE IACC’09, pp. 3697-3701. 2009
Hamid Fadishei, Mehdi Saeedi, Morteza Saheb Zamani, A fast IP routing lookup architecture for multi-gigabit switching routers based on reconfigurable systems, Microprocessors and Microsystems, Volume 32, Issue 4, June 2008, Pages 223-233, ISSN 0141-9331, http://dx.doi.org/10.1016/j.micpro.2008.01.001.
Abstract: With today’s networks complexity, routers in backbone links must be able to handle millions of packets per second on each of their ports. Determining the corresponding output interface for each incoming packet based on its destination address requires a longest matching prefix search on the IP address. Therefore, IP address lookup is one of the most challenging problems for backbone routers. In this paper, an IP routing lookup architecture is proposed which is based on a reconfigurable hardware platform. Experimental results show that the rate of 193 million lookups per second is achieved using our architecture while prefixes can be updated with a rate of 3 million updates per second. Furthermore, it was shown that using our reconfigurable architecture results in rare update failure rate due to resource limitations.
Keywords: IP address lookup; Longest prefix matching (LPM); Reconfigurable hardware; Hashing; Field-programmable gate array (FPGA)
Hamid Fadishei, Morteza Saheb Zamani, Masoud Sabaei, "A novel reconfigurable hardware architecture for IP address lookup," Symposium on Architecture for Networking and Communications Systems, ANCS 2005, pp.81,90, 26-28 Oct. 2005, http://dx.doi.org/10.1145/1095890.1095903
Abstract: IP address lookup is one of the most challenging problems of Internet routers. In this paper, an IP lookup rate of 263 Mlps (Million lookups per second) is achieved using a novel architecture on reconfigurable hardware platform. A partial reconfiguration may be needed for a small fraction of route updates. Prefixes can be added or removed at a rate of 2 million updates per second, including this hardware reconfiguration overhead. A route update may fail due to the physical resource limitations. In this case, which is rare if the architecture is properly configured initially, a full reconfiguration is needed to allocate more resources to the lookup unit.
Keywords: IP address lookup, application specific integrated circuit (ASIC), field-programmable gate array (FPGA), hashing, longest prefix matching, reconfigurable hardware