In order to achieve ultra scalable ip packet switching it is essential to minimize "stopping" of the serial bit streams. In our recent experimental work we demonstrated how this can be achieved with an ultra...
详细信息
In order to achieve ultra scalable ip packet switching it is essential to minimize "stopping" of the serial bit streams. In our recent experimental work we demonstrated how this can be achieved with an ultra-scalable switching architecture reaching multi-terabits per second (10-100 Tb/s) in a single chassis. The implemented testbed uses only off-the-shelf optical and electronic components. The scalability of this architecture is the direct outcome of how global time (i.e., UTC - coordinated universal time) and pipeline forwarding are utilized. The paper presents the design of a prototype switch and experimental activity with it.
In optical networks with sparse wavelength conversion, the optimal converter placement (OCP) problem is one of the most important considerations. In this paper, we propose a heuristic OCP algorithm, called the Optimiz...
详细信息
ISBN:
(纸本)8955191197
In optical networks with sparse wavelength conversion, the optimal converter placement (OCP) problem is one of the most important considerations. In this paper, we propose a heuristic OCP algorithm, called the Optimized Utilization Placement (OUP), to determine proper numbers of converters that should be placed at selected nodes. Performances of the OUP are evaluated by simulations when fixed routing and fixed-alternate routing are used. The simulations justify that the OUP can significantly increase converter gains.
traffic engineering (TE) in multi-region networks is a challenging problem due to the requirement that each region must independently compute its routing decisions based on local observations, yet with the goal of opt...
详细信息
ISBN:
(数字)9781728169927
ISBN:
(纸本)9781728169934
traffic engineering (TE) in multi-region networks is a challenging problem due to the requirement that each region must independently compute its routing decisions based on local observations, yet with the goal of optimizing global TE objectives. Traditional approaches often lack the agility to adapt to changing traffic patterns and thus may suffer hefty performance loss under highly dynamic traffic demands. In this paper, we propose a data-driven framework for multi-region TE problems, which makes novel use of multi-agent deep reinforcement learning. In particular, we propose two reinforcement learning agents for each region, namely T-agents and O-agents, to control the terminal traffic and outgoing traffic, respectively. These distributed agents collect local link utilization statistics within their regions, optimize local routing decisions, and observe the resulting congestion-related reward. To facilitate these agents for optimizing global TE objectives, we tailor the agent design carefully including input, output, and reward functions. The proposed framework is evaluated extensively using real-world network topologies (e.g., Telstra and Google Cloud) and synthetic traffic patterns (e.g., the Gravity model). Numerical results show that comparing with existing protocols and single-agent learning algorithms, our solution can significantly reduce congestion and achieve nearly-optimal performance with both superior scalability and robustness. Throughout our simulations, over 90% of tests limit congestion within 1.2 times the global optimal solution.
Recently, video data transfer is one of the major types of traffic over the Internet. However, it is impossible for the current Internet to guarantee the quality of service (QoS) for video transfer and an influx of a ...
详细信息
Recently, video data transfer is one of the major types of traffic over the Internet. However, it is impossible for the current Internet to guarantee the quality of service (QoS) for video transfer and an influx of a large amount of video data into the Internet may cause serious network congestion. To resolve these problems, we previously proposed a protocol using congestion control based on rate control of the video coding level and data transfer level, and evaluated this protocol using software simulation. However, in order to apply this protocol to real video communication, our previous protocol had some problems, mainly with regard to efficiency. In this paper we present an improved specification of a video transfer protocol with two level rate control. This paper also describes the results of implementation and evaluation of an actual video transfer system using our improved protocol, and a commercial video tool with H.261 encoding. The results show our protocol can transfer video data effectively and fairly through a congested network.
Many schemes to reduce the inter-MAP handoff delay in hierarchical mobile ipv6 have been proposed but the previous schemes waste relatively large network resources to decrease the path rerouting delay. In this paper, ...
详细信息
Many schemes to reduce the inter-MAP handoff delay in hierarchical mobile ipv6 have been proposed but the previous schemes waste relatively large network resources to decrease the path rerouting delay. In this paper, we propose the 2-layered MAP concept, where the seamless inter-MAP handoff can be supported regardless of path rerouting time. As a result, the waste of wired resources and the rate of the inter-MAP handoff can be reduced. From the performance analysis and simulation, the inter-MAP handoff rate for non-real-time traffic is only about 1/3 of the conventional result. Such advantageous features of the proposed scheme neither incur any increase of the total handoff rate nor require additional MAPs.
TT Ethernet is a communication architecture which allows the integration of the standard Ethernet traffic and real-time Ethernet traffic in the same network without invalidating the real-time properties of the real-ti...
详细信息
TT Ethernet is a communication architecture which allows the integration of the standard Ethernet traffic and real-time Ethernet traffic in the same network without invalidating the real-time properties of the real-time traffic. The TT Ethernet switch distinguishes between two classes of traffic. The standard Ethernet traffic is handled in conformance with the existing (standard) Ethernet, whereas the real-time traffic is transmitted with a constant transmission delay. In order to guarantee a constant message transmission delay, the TT Ethernet switch preempts, if necessary, the transmission of standard Ethernet messages, and retransmit the preempted Ethernet message as soon as the transmission of the real-time Ethernet message is finished. The message can be preempted several times before it is successfully transmitted, which decreases the throughput of standard Ethernet messages. In this paper, we propose a segmentation mechanism for standard Ethernet frames in order to increase the throughput. The segmentation mechanism does not change the format of Ethernet frame and is transparent to higher levels of a protocol stack such as TCP/ip.
This paper presents a speed based active queue management scheme. This scheme predicts the incoming and outgoing speed of every active flow in the next sample intervals according to the history data. And then with the...
详细信息
ISBN:
(纸本)0769523331
This paper presents a speed based active queue management scheme. This scheme predicts the incoming and outgoing speed of every active flow in the next sample intervals according to the history data. And then with the queue length of every flow at the end of last a few intervals, predicted incoming and outgoing rate and expected queue length, drop probability in this interval is calculated. The goal of this scheme is to achieve both stabilized queue length and per-flow fairness. Through ns-2 simulation, we compare SBAQM against FRED and SRED and obtain satisfying results.
Internet of Things (IoT) devices in industrial environments can be a significant increase in the network load of the local ip backbone, involving the network performance. To deal with the important amount of network t...
详细信息
Internet of Things (IoT) devices in industrial environments can be a significant increase in the network load of the local ip backbone, involving the network performance. To deal with the important amount of network traffic generated by the IoT devices, this paper proposes the use of IoT gateways in industrial environments (but not limited to) that unifies the data structures improving the size efficiency, and filters the data to be send across the shared ip backbone. To implement this solution some data structures and a multidimensional deadband scheme are presented. Simulations are performed using Networked control Systems as IoT devices to show the feasibility of the proposed solution.
In this paper, the authors develop a novel wireless LAN base station with Web-QoE based QoS control and evaluate its performance by experiment. By adapting a statistical characteristic of Web service traffic to QoS co...
详细信息
In this paper, the authors develop a novel wireless LAN base station with Web-QoE based QoS control and evaluate its performance by experiment. By adapting a statistical characteristic of Web service traffic to QoS control, the base station can keep high QoE without worsening of congestion over wireless LAN; it is implemented by adopting two technologies: OpenFlow and WMM. In the experiment, the authors treat three actual Web services and assess QoE for them. From the experimental results, the authors show the effectiveness of the developed base station.
The use of ABR service on Internet backbones has been recently proposed as a means to improve statistical multiplexing and to recover the bandwidth left over by VBR traffic. While this approach permits far better util...
详细信息
The use of ABR service on Internet backbones has been recently proposed as a means to improve statistical multiplexing and to recover the bandwidth left over by VBR traffic. While this approach permits far better utilization of ATM resources than it is possible otherwise with the CBR service, full link utilization is seldom achieved. We address the reasons why the ABR flow control cannot track the dynamic variations on the available bandwidth along backbone links and we propose an approach to improve resource utilization that relies on bandwidth overbooking. The proposed scheme avoids cell losses and unnecessary rate reductions.
暂无评论