Papers by Mahmoud naghibzadeh
TCP Traffic Classification Using Relaxed Constraints Support Vector Machines
Integration of Practice-Oriented Knowledge Technology: Trends and Prospectives, 2013
ABSTRACT The traffic classification problem is critical for management, security monitoring, and ... more ABSTRACT The traffic classification problem is critical for management, security monitoring, and traffic engineering in computer networks. It has recently taken into consideration by both network operators and researchers. It allows network operators to predict future traffics and detect anomalous behavior and also allows researchers to create traffic models. In this paper, we use a new architecture of support vector machines, namely relaxed constraints support vector machines (RSVMs), to present a traffic classifier that can achieve a high accuracy without any source or destination address or port information. We just use packet length to predict the application class for each flow. RSVM is an efficient and noise-aware implementation of support vector machines that assigns an importance degree to each training sample in such a manner that noisy samples and outliers are given a less degree of importance. Experimental results with UNIBS and AUCKLAND, two sets of traffic traces coming from different topological points in the Internet, show that the proposed classifier is more reliable and has better accuracy.
GRU: Efficient reachability answering for large graphs using united interval labeling
2013 3rd IEEE International Advance Computing Conference (IACC), 2013
ABSTRACT The issue in reachability problem of graph G = (V, E) is whether there is a path between... more ABSTRACT The issue in reachability problem of graph G = (V, E) is whether there is a path between two given nodes or not. This problem plays a key role in areas such as Bioinformatics, Semantic Web, Computer Networks and Social Networks, which have very large graph-structured data. Also, the reachability problem is employed considerably in the graph management and graph algorithms. In this paper, we propose a novel labeling approach for large directed graphs. Our presented method is called GRU (Graph Reachability indexing using United intervals), that can answer reachability queries in constant time even for large graphs. The significant point in this approach is that all the reachability information is computed after indexing time. In addition, this computation is performed only with one time DFS (post-order) traverse and labels are calculated precisely and stored in an efficient way. Analytical and experimental results reveal that effectiveness of our method is more than other interval labeling methods. Furthermore, our approach results show improvement in query time in comparison with GRAIL, which is only a scalable index for reachability queries.
IEEE Transactions on Services Computing
Scheduling workflows in cloud environments is an important issue that many types of research have... more Scheduling workflows in cloud environments is an important issue that many types of research have been conducted in this field. However, these approaches often focus on single workflow scheduling while the need for scheduling multiple workflows is growing. This study aims at presenting a cloud Broker for executing Deadline-constrained Periodic scientific Workflows (BDPW). BDPW acts as a Workflow as a Service (WaaS) broker and uses both reserved and on-demand resources in order to minimize the monetary cost of renting resources from a cloud provider. Furthermore, BDPW uses container technology by executing multiple containerized tasks on the same Virtual Machine (VM) to decrease the provisioning delay of VMs. The proposed broker uses a hybrid scheduling method, i.e., static planning and dynamic scheduling. The static planner uses Resource Leveling Problem (RLP) to provide a scheduling plan and also recognizes the number of reserved resources that should be leased from a provider. Then, the dynamic scheduler tries to assign tasks to the reserved resources based on the primary static plan and leases on-demand instances if necessary. Also, it may make changes to the primary plan due to uncertainties in the task runtimes. The experimental results in CloudSim show that BDPW outperforms baseline algorithms in terms of monetary cost.
Journal of Cloud Computing, May 26, 2022
A workflow is an effective way for modeling complex applications and serves as a means for scient... more A workflow is an effective way for modeling complex applications and serves as a means for scientists and researchers to better understand the details of applications. Cloud computing enables the running of workflow applications on many types of computational resources which become available on-demand. As one of the most important aspects of cloud computing, workflow scheduling needs to be performed efficiently to optimize resources. Due to the existence of various resource types at different prices, workflow scheduling has evolved into an even more challenging problem on cloud computing. The present paper proposes a workflow scheduling algorithm in the cloud to minimize the execution cost of the deadline-constrained workflow. The proposed method, EDQWS, extends the current authors' previous study (DQWS) and is a two-step scheduler based on divide and conquer. In the first step, the workflow is divided into sub-workflows by defining, scheduling, and removing a critical path from the workflow, similar to DQWS. The process continues until only chain-structured sub-workflows, called linear graphs, remain. In the second step which is linear graph scheduling, a new merging algorithm is proposed that combines the resulting linear graphs so as to reduce the number of used instances and minimize the overall execution cost. In addition, the current work introduces a scoring function to select the most efficient instances for scheduling the linear graphs. Experiments show that EDQWS outperforms its competitors, both in terms of minimizing the monetary costs of executing scheduled workflows and meeting user-defined deadlines. Furthermore, in more than 50% of the examined workflow samples, EDQWS succeeds in reducing the number of resource instances compared to the previously introduced DQWS method.
Conceptual summarization using ontologies and nearest neighborhood clustering
... Elahe Gavagsaz, Mahmoud Naghibzadeh, Mehrdad Jalali Department of Software Engineering, Mashh... more ... Elahe Gavagsaz, Mahmoud Naghibzadeh, Mehrdad Jalali Department of Software Engineering, Mashhad Branch, Islamic Azad University, Mashhad, Iran egavagsaz@mshdiau ... over all well-formed terms of the language L [3],[4], [8]. Therefore, the general ontology O = (L, ≤, R ...
A load scheduler for SIP proxy servers: design, implementation and evaluation of a history weighted window approach
International Journal of Communication Systems, 2015
ABSTRACT The widespread use of Session Initiation Protocol as a signalling protocol has created v... more ABSTRACT The widespread use of Session Initiation Protocol as a signalling protocol has created various challenges. An important one is that its throughput can be severely degraded when an overload happens in the proxy server because of several retransmissions from the user agent. One common approach to overcome this problem is ‘load balancing’. A balancer needs to know the status of proxy servers, which are continuously gathered implicitly or explicitly. Implicit methods have averagely less overhead than explicit ones. This paper attempts to prevent throughput reduction by balancing the loads among available proxy servers properly using an implicit mechanism called History Weighted Average Response time. The proposed algorithm is robust because it incurs no extra processing to proxy servers. The novelty of the mechanism is making use of ‘response time history’ to estimate the load being currently processed on servers. By implementing in a real testbed, throughput and scalability are improved compared with an important state-of-the-art similar algorithm. This improvement stems from no need for modification in SIP protocol, easy implementation and application, simple computations for making decision and no need for extra feedback between servers and load balancer. Copyright © 2015 John Wiley & Sons, Ltd.
Paper: MODELING AND PERFORMANCE EVALUATION OF DISTRIBUTED SYSTEMS WITH COORDINATOR
Paper: PARALLEL ALGORITHMS FOR AUTOMATIC DATABASE NORMALIZATION
A message passing protocol for small-scale distributed real-time systems
World Automation Congress, 2004
In distributed real-time systems, every node is usually responsible for performing pre-assigned l... more In distributed real-time systems, every node is usually responsible for performing pre-assigned local activities that are required for its physical environment. Nodes are connected via high-speed communication links and information flow is made possible by using message-passing techniques. Reliability, in-time delivery of messages, priority support, and fault-tolerance are amongst the requirements for such systems. Current Techniques lack all or most of these requirements. This is because they are general-purpose techniques and not tailored to match specific requirements needed by small-scale distributed real-time systems. In this piper, the Round Data Mailer technique is further studied and a layered protocol is proposed.
A fault-tolerant mutual exclusion resource reservation protocol for clustered mobile Ad hoc networks
Page 1. A Fault-Tolerant Mutual Exclusion Resource Reservation Protocol for Clustered Mobile Ad h... more Page 1. A Fault-Tolerant Mutual Exclusion Resource Reservation Protocol for Clustered Mobile Ad hoc Networks Mohammad Moallemi, Mohammad Hossien Yaghmaee Moghaddam, Mahmoud Naghibzadeh Department of ...
Quantitative comparison of job-level dynamic scheduling policies in parallel real-time systems
Abstract-Scheduling algorithms play an important role in design of real-time systems. Due to high... more Abstract-Scheduling algorithms play an important role in design of real-time systems. Due to high processing power and low price of multiprocessors, real-time scheduling in such systems is more interesting; however, more complicated. Earliest Deadline First (EDF) and Least ...
Delayed rate monotonic with semi-partitioned technique adapted to multiprocessors
ABSTRACT This paper presents a semi-partitioned scheduling algorithm for periodic tasks on multip... more ABSTRACT This paper presents a semi-partitioned scheduling algorithm for periodic tasks on multiprocessor platforms. The scheduling policy used within each processor is based on delayed rate monotonic algorithm. Delayed rate monotonic is a modified version of rate monotonic algorithm, which can achieve higher processor utilization than rate monotonic. First, it is formally proved that any task which is feasible under rate monotonic is feasible under delayed rate monotonic, as well. Then, based on delayed rate monotonic, a new scheduling algorithm called adapted delayed rate monotonic (ADRM) is designed for multiprocessor systems. ADRM is composed of delayed rate monotonic algorithm and a semi-partitioned technique. Semi-partitioned is a new approach for allocating tasks on multiprocessor platforms. According to the simulation results, ADRM is safer against possible overload than rate monotonic algorithm. Then, a procedure is proposed to achieve processor utilization under ADRM algorithm more than the rate monotonic algorithm.
Power management in large-scale computational environments can significantly benefit from predict... more 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 machineindependent workload characteristics and workloadindependent 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.
MMUF: An Optimized Scheduling Algorithm for Dynamically Reconfigurable Real-Time Systems
2006 2nd International Conference on Information & Communication Technologies
In this paper we compare our proposed hybrid scheduling algorithm which is a modification of Maxi... more In this paper we compare our proposed hybrid scheduling algorithm which is a modification of Maximum Urgency First (MUF) scheduling algorithm with MUF scheduling algorithm. The maximum urgency first algorithm combines the advantages of fixed and dynamic scheduling to provide the dynamically changing systems with flexible scheduling. This algorithm, however, has a major shortcoming due to its scheduling mechanism which
The high cost of supercomputers and the need for large-scale computational resources has led to t... more The high cost of supercomputers and the need for large-scale computational resources has led to the development of network of computational resources known as Grid. To better use tremendous capabilities of this large scale distributed system, effective and efficient scheduling algorithms are needed. Many such algorithms have been designed and implemented. We introduce a new scheduling algorithm based on two conventional scheduling algorithms, Min-Min and Max-Min, to use their cons and at the same time, overcome their pros. This heuristic scheduling algorithm, called min-min min-max selective, is evaluated using a grid simulator called GridSim by comparing to its performance against the two basic heuristics which it comes from. The simulation results show that the new heuristic can lead to significant performance gain for a variety of scenarios.
RADMA: A Mobile Agent-Based Resource Discovery Approach For the Grid Environment
Nexus 2007���3D Soccer Simulation Team Description
The yielding-first rate-monotonic scheduling approach and its efficiency assessment
In real-time (RT) computing systems, a service function execution including an output action is t... more In real-time (RT) computing systems, a service function execution including an output action is typically subject to a deadline. Scheduling algorithms play a significant role in the design of such systems. They are the major factor in determining the execution safety, i.e., the possibility of always executing control service functions without missing deadlines in a given execution engine possessing a limited set of execution resources. The rate-monotonic (RM) fixed-priority scheduling approach is one of the oldest scheduling approaches devised for use in RT computing systems [1–3]. It can be used effectively in a narrow class of RT application situations. In this paper, we attempt to add some to the knowledge-base related to the RM approach. The essence of the RM approach is: (E1) To structure the RT application software in the form of a set of cyclic fixed-priority processes, each iterating the sequence of receiving a service request signal coming from a controlled device or timer and executing a control service function; (E2) To make the priority number associated with each process to be either inversely proportional to the minimum request interval (MRI), i.e., the minimum interval between two successive arrivals of service request signals, or proportional to the maximum request rate which is the maximum rate of arrivals of service request signals, which is also called the maximum service rate; and (E3) To select the highest-priority process at every selection point.
The Impact of Refactoring Based on Naming Pattern Detection on RiMOM Result
Journal of Systems Architecture, 1996
Uploads
Papers by Mahmoud naghibzadeh