Sharif Digital Repository

In nanometer technologies, the effect of process variation is observed in Multi-Processor System on Chip (MPSoC) in terms of variation in processors‟ frequency and leakage power. Traditionally, only worst case values of the system parameters were concerned and a worst-case optimization algorithm was employed for an application under design. As previous researches have shown these algorithms are not optimal in terms of parametric yield compared with newly employed statistical optimization algorithms. In this project, we have considered the problem of simultaneously selecting MPSoC architecture (which includes type and number of processors and the communication media) and task and... 

Process Variation is seen as statistical variations in leakage current and delay of transistors in nano-scale technologies. The amount of process variations increase as the size of transistors decrease by technology scaling such that those effects can be seen in frequency of MPSoC (Multi-Processor System-on-Chip) cores and their leakage power deviation. These variations cause the tasks duration and power consumption fluctuate in different processors in an MPSoC instance. Consequently, some chip instances of the same MPSoC may consume more time and power than their considered limitations. Hence considering the process variation is necessary and required for MPSoC optimization at different... 

Size and number of high-performance data centers are fast growing all over the world in recent years. The growth in the leakage power consumption of servers along with its exponential dependency on the ever increasing process variation in nanometer technologies have made it inevitable to move toward variation-aware power reduction strategies. In this thesis, we simultaneously apply thermal- and variation-aware server placement and chassis consolidation methods to reduce total power consumption of data centers. We introduce two server placement heuristics as well as an Integer Linear Programming (ILP)-based server placement method based on power consumption of each server and the data center... 

With the proliferation of Internet-based applications and malicious attacks, security has become one of the most influential aspects in the network and, it should be considered from the beginning steps of designing the network infrastructure. Based on the fact that pattern matching is considered as one of the most important roles of security devices or applications, it becomes an important procedure in firewalls that have been classified as security equipments which adopt a security mechanism in order to restrict the traffic exchanged between networks and particular users or certain applications. While the trend of using compressed traffic is drastically increasing, this type of traffic is... 

Mobile payment has already become an integral part of modern life for daily banking needs. Many payment systems are now providing financial services to customers. Most of these systems need and rely on online network access to provide their services. On the other hand, there are many regions in the world that have no network access or the network quality is poor with intermittent failures. Customers in these areas are unable to transfer money or buy goods and pay for it using their mobile devices by such above systems. This motivated us to find a way for transferring money without need for network availability. As a result we devised an offline peer-to peer mobile payment system (MobiPOS).... 

In the past few years, because of the high energy price in data centers, optimizing energy consumption has become of real importance. The disproportionate energy consumption of network devices to their utilization has led them to become one of the biggest consumers of energy inside data centers. On the other hand, the stochastic nature of network traffic among virtual machines has made the assumption of known or fixed communication volume superficial. In this research we considered a stochastic distribution of network traffic and communication volume among virtual machines trying to change the communication inside data center to communication inside hosts. Initially, we start by evaluating... 

Recently, The Earth has become warm rapidly over the past few years. This global warming is not only caused by natural phenomena, and further increased by humans and the amount of greenhouse gases in the atmosphere. Around 40 to 50 percent of energy consumption is related to information technology and communications. However, the rapid growth of electricity consumption in data centers, in addition to increased costs, data center challenges faced serious environmental pollution and more greenhouse gases. Use of renewable energy and reducing electricity costs in data centers are the important subjects for researchers around the world. Many researchers provide solutions to reduce environmental... 

Cloud computing is a way to increase capacity or add capabilities dynamically without investing in new infrastructure. The purpose of this project is the use of specialized algorithms for efficient energy management for cloud computing environments.One of the issues that is very important with respect to cloud computing is considering lifetime of servers or physical machines. Servers are some of the most important and critical elements of cloud computing. Server costseffect on the costs of the entire cloud computing system and the interference of serverseffects on entire system. Lifetime of each sever is dependent on a few factors, the most important of which is the number of switching... 

The ever increasing density and performance of FPGAs, has increased the importance and popularity of soft processors. One major research concern in this regard lays in the field of energy efficiency of the system on FPGA. This work is particularly focused on the energy efficiency of multiprocessor structures on FPGA using system level techniques. The growing gap between the speed of processors and memories can partly be compensated through memory hierarchy, i.e. caches. Since memory accesses follow a non-uniform distribution, and vary from one application to another, variable set-associative cache architectures have emerged. In this thesis, two novel cache architecture, primarily aimed at... 

Using accelerators inside high speed servers can reduce execution time of applications and total power usage of the system. Sharing accelerator between virtual machines of a server decrease both cost and power, however it won’t provide the gained speedup of using dedicated accelerator for each virtual machines. Creation of an appropriate set of accelerators required for virtual machines, management of accesses to the accelerator, prioritizing and scheduling of requests and reconfiguration type of accelerator are the most important challenges that this project has been dealt with. The main objective of this project is implementing the necessary infrastructure to share an FPGA-based... 

Today, we have witnessed many changes in computer science. Rapid and exponential growth of data volumes, increasing a variety of data structures and forms, increasing data generation ratio are some important revolutions in our digital world. Because of these changes, our traditional computing models and processing methods cannot solve new big and sophisticate problems. Big data concepts and big data processing methods come to help us for solving these problems. A lot of parallel and distributed processing platforms built for the applications execution on a huge volume of data. Apache spark and Hadoop are two powerful platforms that each has some advantage and disadvantage. With growth... 

In recent years, big data processing plays an important role in the era of information technology. The exponential growth of big data volume increases the need for data centers and infrastructures with more processing power. Due to dark silicon and scalability limitations in deep-submicron, the increasing trend of server performance slows down. Therefore, hardware accelerators such as FPGA and GPU are become increasingly popular for improvement of data center processing power. There are two types of big data processing based on the application: stream processing and batch processing. With the widespread use of social networks, online control systems and internet of things services, the... 

Distributed stream processing frameworks (DSPFs) are used for real-time processing of big data. Apache Storm is one of the most popular stream processing systems in industry today. By increasing data generation rate we need new methods to overcome processing requirements of DSPFs like Apache Storm. In this thesis we investigate the feasibility of incorporation FPGA acceleration into Apache Storm. Using FPGAs as co-processors in powerful servers can improve performance and accelerate processing of streaming data by increasing parallelism, decreasing processing time of each processing units and decreasing communication delay between these units. Our design includes a hardware part that... 

One of the most important categories of big data processing is stream processing. In stream processing, processing of data is performed simultaneously with the production of data. one of the most well-known frameworks used for stream processing is Apache Storm. By default, Storm uses a round-robin scheduler to allocate tasks to physical machines. This scheduler randomly performs scheduling and assignment of tasks to physical machines without considering the processing power of physical machines and processing tasks, which makes it impossible to properly utilize the processing resources. In this paper, a scheduling algorithm and resource allocation have been proposed based on the processing... 

Nowdays a large number of companies are faced with Big Data processing. Due to the lack of budget or time, it may not be possible to process all the input data. Data variety is one of the main features of Big Data. Considering Data variety can help us to solve the problem of resource capacity. In this research, we focused on the impact of data variety on the performance of Big Data processing. In the first part, we offer a solution for increasing the performance of progressive Big Data processing. We provide a simple low-overhead mechanism to quickly assess the significance of each data portion, and show its effectiveness in finding the best ranking of data portions. We continue by... 

Today, due to the increasing use of fossil fuel-powered vehicles, attention has been paid to advanced societies by the amount of pollutants emitted by these vehicles and the need to control and reduce them. As the number of cars increases, the volume of data for processing increases, which increases the need for data centers and processing infrastructure with more computing power. Therefore, in this research, we intend to implement a program that continuously categorizes cars according to specific criteria in terms of emission levels. For this purpose we used Apache Storm, a data stream processing framework, to implement the program.In this research, we develop a simulation for cars with the... 

In recent years, and with the development of information technology in various aspects of life, the growth of energy consumption in this sector has accelerated. With the advent of cloud computing and the development of its users, data centers have become one of the largest consumers of energy in the world. Adverse effects of fossil fuel consumption on the one hand and the global energy crisis as a limiting factor for development, on the other hand, have doubled the importance of managing the issue and reducing energy consumption in this sector. The two main energy-related concerns in data center construction are clean energy and energy management. The use of new energy and natural facilities... 

Nowadays, one of the challenges in the field of bioinformatics is the excess processed data volume such that this data volume resulted from a complete genome sequence of a species can be up to hundreds gigabytes. Every time that we talk about increasing data volume, data storage, transforming, and the process will become of interest. Moreover, considering the presence of portable sequencer devices in the market and the limitations of process outside of the lab environments, this problem becomes of more critical importance. Fortunately, due to the nature of the genome data and their redundancy, specific algorithms to compress them have been introduced to the market. In this thesis, we chose... 

Recently, portable genome sequencing devices have been introduced to the market, which have also made it possible to provide these services in remote locations or outside the laboratory. The amount of raw data from the readings of a sequencer for the entire genome of a human or plant can be in the hundreds of gigabytes, making it difficult and expensive to maintain and transfer to the center for such sequencing. Fortunately, these readings have a lot of redundancy, and many new algorithms have been proposed to compress them based on the intrinsic properties of this data. Sequencing devices were mainly used in the laboratory environment, which naturally had virtually unlimited access to urban... 

Automatic Speech Recognition (ASR) systems are a significant part of Personal Assistants in mobile phones. But because of the time-dependent nature of ASR systems, they are computation and memory-intensive tasks. On the other hand, mobile devices utilize a Low-Power design to extend battery life and improve user experience, making them incompatible with heavy-loaded tasks such as ASR systems. For instance, if we run an inference with a 60 seconds audio file on a well-known open-sourced Speech Recognition System named DeepSpeech, it will only take 49 seconds for a desktop PC to generate the results. Still, a mobile phone with ARM64 architecture with the same input file will take 92 seconds to...