-
Extending the Control Plane of Container Orchestrators for I/O Virtualization
Authors:
Garegin Grigoryan,
Minseok Kwon,
M. Mustafa Rafique
Abstract:
Single Root Input/Output Virtualization (SR-IOV) is a standard technology for forking a single PCI express device and providing it to applications while ensuring performance isolation. It enables container orchestrators to share a limited number of physical network interfaces without incurring significant virtualization overhead. The allocation of virtualized network devices to containers, however…
▽ More
Single Root Input/Output Virtualization (SR-IOV) is a standard technology for forking a single PCI express device and providing it to applications while ensuring performance isolation. It enables container orchestrators to share a limited number of physical network interfaces without incurring significant virtualization overhead. The allocation of virtualized network devices to containers, however, needs to be more configurable based on the bandwidth needs of running applications. Moreover, container orchestrators' network control over the virtualized interfaces is limited by the abilities of SR-IOV. We explore the design considerations for a system with controlled SR-IOV virtualization and present ConRDMA, a novel architecture that enables fine control of RDMA virtualization for containers. Our evaluation shows that ConRDMA enables containers to use RDMA allocated bandwidth more efficiently and to select best-suited nodes to meet their varying communication requirements.
△ Less
Submitted 9 May, 2025;
originally announced May 2025.
-
Suitability of FPS and DPS in NOMA for Real-Time and Non-Real Time Applications
Authors:
Moontasir Rafique,
Abdullah Alavi,
Aadnan Farhad,
Mohammad T. Kawser
Abstract:
Non-Orthogonal Multiple Access (NOMA) is a popular solution for supporting a high number of users and along with significant bandwidth in 5G cellular communication. By using a technique called cooperative relaying, the same data is sent to all the users, and one user can relay data to the other. In order to provide enough power for the users, energy harvesting techniques have been introduced with…
▽ More
Non-Orthogonal Multiple Access (NOMA) is a popular solution for supporting a high number of users and along with significant bandwidth in 5G cellular communication. By using a technique called cooperative relaying, the same data is sent to all the users, and one user can relay data to the other. In order to provide enough power for the users, energy harvesting techniques have been introduced with Simultaneous Wireless Information and Power Transfer (SWIPT) coming to prominence in recent times. In this paper, analysis has been made comparing two different power allocation schemes in NOMA, Fixed Power allocation Scheme (FPS) and Dynamic Power allocation Scheme (DPS). The comparisons were made in terms of their performance and characteristics while undergoing SWIPT. It has been found that by using DPS, an almost 25% increase in peak spectral efficiency can be obtained compared to FPS. However, DPS suffers from a higher outage probability as the increase of power causes the signal bandwidth to drop below the target rate a significant number of times. Based on the detailed results, conclusions were drawn as to which power allocation coefficient scheme would be used in real-time and non-real time communication standards, respectively. The results suggest that for real-time communication, FPS is more suitable while for non-real-time communication, DPS appears to work better than FPS.
△ Less
Submitted 23 August, 2021;
originally announced August 2021.
-
Downlink Performance Enhancement of High Velocity Users in 5G Networks by Configuring Antenna System
Authors:
Mariea Sharaf Anzum,
Moontasir Rafique,
Md. Asif Ishrak Sarder,
Fehima Tajrian,
Abdullah Bin Shams
Abstract:
A limitation of bandwidth in the wireless network and the exponential rise in the high data rate requirement prompted the development of Massive Multiple-Input-Multiple-Output (MIMO) technique in 5G. Using this method the ever rising data rate can be met with the increment of the number of antennas. This comes at the price of energy consumption of higher amount, complex network setups and maintena…
▽ More
A limitation of bandwidth in the wireless network and the exponential rise in the high data rate requirement prompted the development of Massive Multiple-Input-Multiple-Output (MIMO) technique in 5G. Using this method the ever rising data rate can be met with the increment of the number of antennas. This comes at the price of energy consumption of higher amount, complex network setups and maintenance. Moreover a high-velocity user experiences unpredictable fluctuations in the channel condition that deteriorates the downlink performance. Therefore a proper number of antenna selection is of paramount importance. This issue has been addressed using different categories of algorithms but only for static users. In this study we proffer to implement antenna diversity in closed loop spatial multiplexing MIMO transmission scheme by operating more number of reception antennas than the number of transmission antennas for ameliorating the downlink performance of high-velocity users in case of single user MIMO technology. In general our results can be interpreted for large scale antenna systems like Massive MIMO even though a 4x4 MIMO system has been executed to carry out this study here. Additionally it shows great prospects for solving practical life problems like low data rate and call drops during handover to be experienced by cellular users traveling by high-speed transportation systems like Dhaka Metro Rail. The cell edge users are anticipated to get benefits from this method in case of SU-MIMO technology. The proposed method is expected to be easily implemented in the existing network structures with nominal difficulties.
△ Less
Submitted 10 August, 2021;
originally announced August 2021.
-
Configuring Antenna System to Enhance the Downlink Performance of High Velocity Users in 5G MU-MIMO Networks
Authors:
Md. Asif Ishrak Sarder,
Fehima Tajrian,
Moontasir Rafique,
Mariea Anzum,
Abdullah Bin Shams
Abstract:
An exponential increase in the data rate demand prompted several technical innovations. Multi User Multiple Input Multiple Output (MU-MIMO) is one of the most promising schemes. This has been evolved into Massive MIMO technology in 5G to further stretch the network throughput. Massive MIMO tackles the rising data rate with the increase in the number of antenna. This comes at the price of a higher…
▽ More
An exponential increase in the data rate demand prompted several technical innovations. Multi User Multiple Input Multiple Output (MU-MIMO) is one of the most promising schemes. This has been evolved into Massive MIMO technology in 5G to further stretch the network throughput. Massive MIMO tackles the rising data rate with the increase in the number of antenna. This comes at the price of a higher energy consumption. Moreover the high velocity users in MU-MIMO scheme experiences a frequent unpredictable change in the channel condition that degrade its downlink performance. Therefore a proper number of antenna selection is of paramount importance. This issue has been addressed using machine learning techniques and Channel State Information (CSI) but only for static users. In this study we propose to introduce antenna diversity in spatial multiplexing MU-MIMO transmission scheme by operating more number of reception antenna compare to the number of transmission antenna. The diversity improves the downlink performance of high velocity users. In general our results can be interpreted for large scale antenna systems like Massive MIMO. The proposed method can be easily implemented in the existing network architectures with minimal complexity. Also it has the potential for solving real-life problems like call drops and low data rate to be experienced by cellular users traveling through high-speed transportation systems like Dhaka MRT project
△ Less
Submitted 5 August, 2021;
originally announced August 2021.
-
Impact of Receiver Antenna Polarization and Resource Scheduler on the Downlink Performance of High Velocity Users in 5G Millimeter Wave Small Cell Technology
Authors:
Fehima Tajrian,
Md. Asif Ishrak Sarder,
Mariea Sharaf Anzum,
Moontasir Rafique,
Abdullah Bin Shams
Abstract:
With the occupancy of the existing frequency spectrum the demands of the skyrocketing data traffic paved the path for 5G millimeter Wave (mmWave) technology. The wide frequency spectrum and high directivity can elevate the mean data rate. This performance gain comes at the cost of a higher pathloss which limits the use of mmWave technology to small cells. Also at such high frequencies the multipat…
▽ More
With the occupancy of the existing frequency spectrum the demands of the skyrocketing data traffic paved the path for 5G millimeter Wave (mmWave) technology. The wide frequency spectrum and high directivity can elevate the mean data rate. This performance gain comes at the cost of a higher pathloss which limits the use of mmWave technology to small cells. Also at such high frequencies the multipath components and blockages appear as persistent barriers leading to alteration of antenna polarization, affecting resource scheduler performance and throughput degradation. The impact of these barriers for high velocity user equipments (UEs) in mmWave network is still left to be studied. Using closed loop spatial multiplexing transmission scheme this study analyses the effect of receiver antenna polarization and types of resource scheduler used in base station on the downlink performance of mobile users (0-120kmph) in the mmWave small cell network. Thorough investigation has been conducted to infer which antenna combination should be advantageous under different scheduling algorithms for high mobility UEs. Our results indicate that the scheduler performance is complimented by the receiver antenna polarization, and by appropriate selection a better downlink performance can be sustained. Also reception under linear polarization performs better over circular polarization for high velocity UEs.
△ Less
Submitted 11 September, 2021; v1 submitted 5 August, 2021;
originally announced August 2021.
-
N 2 C : Neural Network Controller Design Using Behavioral Cloning
Authors:
Shoaib Azam,
Farzeen Munir,
Muhammad Aasim Rafique,
Ahmad Muqeem Sheri,
Muhammad Ishfaq Hussain,
Moongu Jeon
Abstract:
Modern vehicles communicate data to and from sensors, actuators, and electronic control units (ECUs) using Controller Area Network (CAN) bus, which operates on differential signaling. An autonomous ECU responsible for the execution of decision commands to an autonomous vehicle is developed by assimilating the information from the CAN bus. The conventional way of parsing the decision commands is mo…
▽ More
Modern vehicles communicate data to and from sensors, actuators, and electronic control units (ECUs) using Controller Area Network (CAN) bus, which operates on differential signaling. An autonomous ECU responsible for the execution of decision commands to an autonomous vehicle is developed by assimilating the information from the CAN bus. The conventional way of parsing the decision commands is motion planning, which uses a path tracking algorithm to evaluate the decision commands. This study focuses on designing a robust controller using behavioral cloning and motion planning of autonomous vehicle using a deep learning framework. In the first part of this study, we explore the pipeline of parsing decision commands from the path tracking algorithm to the controller and proposed a neural network-based controller ($N^2C$) using behavioral cloning. The proposed network predicts throttle, brake, and torque when trained with the manual driving data acquired from the CAN bus. The efficacy of the proposed method is demonstrated by comparing the accuracy with the Proportional-Derivative-Integral (PID) controller in conjunction with the path tracking algorithm (pure pursuit and model predictive control based path follower). The second part of this study complements $N^2C$, in which an end-to-end neural network for predicting the speed and steering angle is proposed with image data as an input. The performance of the proposed frameworks are evaluated in real-time and on the Udacity dataset, showing better metric scores in the former and reliable prediction in the later case when compared with the state-of-the-art methods.
△ Less
Submitted 30 November, 2021; v1 submitted 1 June, 2020;
originally announced June 2020.
