Kubernetes is a powerful tool to manage containerized applications, which is also regarded as one promising platform to support microservices in edge computing. The scheduler is a key component of Kubernetes. It allocates each pod (i.e., a set of running containers) to one worker node (i.e., a machine). The default scheduler in Kubernetes is designed for the cloud environment containing homogeneous nodes. However, IoT edge nodes usually have various computing power and network bandwidth. This article proposes a delay-aware container scheduling (DACS) algorithm to address the issue of node heterogeneity in edge computing. To efficiently assign pods to worker nodes, DACS takes account of not only residual resources of worker nodes but also potential delays caused by the pod assignment. We build a Kubernetes cluster by VMware to evaluate system performance. Experimental results reveal that DACS can significantly reduce both processing and network delays, thereby helping Kubernetes perform more efficiently in an edge environment.

Software-defined networking (SDN) makes network management easier by using a controller to govern all switches, but the controller may become a performance bottleneck. Distributed SDN control is a promising solution, which lets multiple controllers divide the work, where each controller manages a part of the network. Switch migration is one common means to the load balance of controllers, which transfers some switches to different subnets based on the workloads of their controllers. The paper proposes a time-sharing switch migration (TSSM) scheme to provide more refined load sharing for controllers, which allows two controllers to share a switch’s load sequentially in the same period. When a controller is overloaded, TSSM finds assistant controllers to share its workload by selecting proper switches for migration and also deciding the time to perform migration. In this way, the workload of each controller can be kept below a given threshold. We implement the TSSM scheme on the open network operating system (ONOS) to attest to its feasibility. Experimental results show that TSSM can reduce 98% of the occurrences of overload for controllers as compared with the original OpenFlow method. Moreover, TSSM can save about 78% of the migration cost than the churn-triggered migration method.

Internet-of-Things (IoT) devices are widely deployed and many of them have limited capabilities. Constrained application protocol (CoAP) is developed to integrate such devices into the Web environment. Specifically, client devices register their interested data with IoT devices, which will keep notifying them of the status of sensing data. However, client devices have different demands in terms of minimum and maximum periods to receive notifications, whereas IoT devices usually offer notifications at regular intervals. Some IoT devices would send many more notifications than necessary, which wastes bandwidth and energy. To conquer this problem, this article proposes a group-based message management (GMM) framework for CoAP proxies to efficiently coordinate the sending of notifications from IoT devices and forward them to client devices. GMM curtails superfluous notifications by three modules. The scheduling module groups client devices based on their demands and finds an optimal observation period for each group. The caching module adjusts the max-age value to allow a proxy reusing its cached notifications to answer requests. The merging module combines the notifications originated from different IoT devices queried by the same client device so as to save bandwidth. The simulation results show that GMM not only reduces unnecessary notifications but also conserves the energy of devices.

Densely deploying small cells will be a solution to provide explosive data requirements in fifth-generation networks. Because users often cluster together in popular locations of an urban area, congested small cells in the popular sites are also gathered. Traditional load balancing schemes generally only consider neighboring cells when offloading users are not suitable for clusters of overloaded cells. This paper considers groups of overloaded cells in the load balancing problem. The objective is to maximize the quality-of-service satisfaction ratio. To solve the problem, we propose a multidepth offloading algorithm with the consideration of the radio resource allocation. The proposed multidepth offloading algorithm can be applied no matter that congested small cells are gathered or not. Compared with a previous offloading algorithm and a baseline, the simulation results show that our proposed algorithm can increase 16% QoS satisfaction ratio in a real user distribution and 13% QoS satisfaction ratio in a clustered user placement.

In the fifth-generation mobile network (5G), exponential growth in bandwidth requirements and massive connectivity have become new challenges. Ultra-dense network deployment is a promising technology to meet the needs of the future network. However, ultra-dense network deployment also caused significant interference between cells. Coordinated multipoint is a technology proposed by 3GPP to reduce interference between cells effectively. To this end, we propose a novel method for user association and resource allocation based on coordinated multipoint. First, a cell-filtering method is proposed to preserve the high resource efficiency connection. Second, location-based and load-based clustering is used in our method to reduce network complexity. Depend on the results of clustering, a competition-based resource allocation scheme is proposed for intra-cluster and inter-cluster resource allocation. At the end of our method, connection tuning algorithm is performed to maximize the number of users who meet their Qu's requirements through resource reallocation. The goal of our proposed method is to maximize the number of users who meet their Qu's requirements. Simulation results show that compared with traditional resource allocation methods, the proposed method has better performance of Qu's satisfaction rate.

In vehicular ad hoc networks (VANETs), due to highly mobile and frequently changing topology, available resources and transmission opportunities are restricted. Transmitting frames efficiently within transmission opportunities is very important for system performance. Generally, there are many transmissions between many senders and receivers in a congested environment, which leads to network congestion and reduces transmission efficiency. To address this, we propose a burst transmission and frame aggregation (FAB) scheme to enhance transmission opportunity (TXOP) efficiency of IEEE 802.11p. The proposed scheme considers aggregation and TXOP techniques. Such techniques are useful for improving transmission performance. FAB aggregates frames in the relay node and utilizes the TXOP to transmit these frames to the next hop with a burst transmission. FAB also select multiple relay nodes to improve the efficiency of data forwarding if data is large and can not be forwarded by the node with the largest weight value. This FAB scheme increases transmission efficiency within one TXOP while decreasing transmission overheads. FAB can also support traffics with multiple classes. The proposed strategy is evaluated using various metrics such as end-to-end delays, collision rates, and transmission overheads. Simulation results show that the proposed FAB scheme can significantly improve the performance of inter-vehicle communications.

In vehicular ad hoc networks (VANETs), due to highly mobile and frequently changing topology, available resources and transmission opportunities are restricted. Transmitting frames efficiently within transmission opportunities is very important for system performance. Generally, there are many transmissions between many senders and receivers in a congested environment, which leads to network congestion and reduces transmission efficiency. To address this, we propose a burst transmission and frame aggregation (FAB) scheme to enhance transmission opportunity (TXOP) efficiency of IEEE 802.11p. The proposed scheme considers aggregation and TXOP techniques. Such techniques are useful for improving transmission performance. FAB aggregates frames in the relay node and utilizes the TXOP to transmit these frames to the next hop with a burst transmission. FAB also select multiple relay nodes to improve the efficiency of data forwarding if data is large and can not be forwarded by the node with the largest weight value. This FAB scheme increases transmission efficiency within one TXOP while decreasing transmission overheads. FAB can also support traffics with multiple classes. The proposed strategy is evaluated using various metrics such as end-to-end delays, collision rates, and transmission overheads. Simulation results show that the proposed FAB scheme can significantly improve the performance of inter-vehicle communications.

Information and communication technologies have improved the quality of Intelligent Transportation Systems (ITS). The real-time traffic information has traditionally been collected via stationary vehicle detectors or GPS-based probe cars. Compared to the traditional ways, estimating traffic information from Cellular Floating Vehicle Data (CFVD) is more cost-effective, and easier to acquire. In this paper, this study proposes a novel approach to evaluate the relation of call arrival, handover, traffic flow, and traffic density. Moreover, the traffic speed is estimated by the proposed approach according to CFVD. Through the analytical analysis, this study analyzes the effects of traffic information (e.g. traffic flow and vehicle speed) and communication behaviors (e.g. call arrival rate and call holding time) on handovers and call arrivals. In the simulation, this study compares the estimated traffic information with the real traffic information. The experiment results show that the accuracy of traffic speed estimation is 89.75%. Therefore, the proposed approach can be used to estimate traffic speed from CFVD for ITS.

ABSTRACT A hop‐aware and energy‐based buffer management scheme (HEB) is proposed in this paper. HEB can provide better quality of service to packets with real‐time requirements and improve MANET power efficiency. In our algorithm, the buffer is divided into real‐time and non‐real‐time partitions. We consider the number of hops passed, the power levels of the transmitting node, the predicted number of remaining hops, and waiting time in the buffer to determine packet transmission priority. In addition, specialized queue management and a probabilistic scheduling algorithm are proposed to decrease retransmissions caused by packet losses. Mathematical derivations of loss rates and end‐to‐end delays are also proposed. Coincidence between mathematical and simulation results is also shown. Finally, the HEB is compared with first in first out, random early detection, and hop‐aware buffering scheme. Simulation results show that the proposed algorithm reduces loss rates, power consumption, and end‐to‐end delays for real‐time traffic, considerably improving the efficiency of queue management in MANET. Copyright © 2012 John Wiley & Sons, Ltd.

Efficient handover schemes are essential to the success of mobile networking. However, there are time-consuming procedures involved in the process, such as duplicate address detection, packet buffering and tunneling. A novel scheme featuring multicast technique is developed and presented in this article. The proposed approach switches to multicast addressing during handover and switches back to normal unicast addressing after all required operations are done. By switching to multicast addressing, a mobile node can continue the reception of packets from its corresponding node before its new care-of address is actually validated. As a result, transmission disruption can be effectively minimized. In addition, a simple but effective buffer control scheme is designed to reduce the possible packet loss and to prevent the out-of-order problem. Analytical study reveals that improved performance can be guaranteed, as reflected in the simulation results.

Abstract The deployment of directional antennas offers many advantages, such as transmission range extension, co‐channel interference reducing, the increasing of spatial reuse degree, throughput improving of networks, and transmission power saving. Hence, manyad hocMAC protocols with directional antennas have been proposed. However, these protocols do not provide significant improvement of network performance due to the lack of supporting multiple transmissions and receptions simultaneously. With the adaptive beam‐forming system (Smart Antenna), a newly designed MAC protocol (MARS) that enables nodes with multiple transmissions and receptions is proposed. Simulation results show that our MARS do exploit the advantage of space division multiple access. In terms of total number of data forward and complete sessions in bottleneck nodes, MARS achieves three times better than Novel and nine times better than IEEE 802.11. In addition, the end‐to‐end delay keeps very short. Copyright © 2008 John Wiley & Sons, Ltd.

Abstract Due to the mismatch between the transmission range and interference‐vulnerable range, the reception at a receiving node in a wireless network could still be interfered by a sending node beyond the coverage of its clear‐to‐send (CTS). This is the so‐called large interference range problem. It is a common problem in real environments but overlooked by the standards and related studies. The large interference range problem could lead to transmission control protocol (TCP) instability and TCP unfairness, especially in case of multi‐hop transmission. There exist research works devoted to this issue. However, more power consumption was incurred at the same time. For a power‐efficient solution to this problem, a novel scheme, named power control interference avoidance (PCIA), is developed and presented in this article. The proposed scheme decouples control channel from data channel and use different power level in the transmission of control frames and data frames. By doing so, the large interference problem can be effectively avoided, while retaining a high level of power efficiency. The feasibility and effectiveness of the proposed scheme had been validated by an analytic study on a simplified model and a series of simulations. As revealed in the simulation results, the proposed PCIA scheme outperforms conservative CTS reply (CCR) and IEEE 802.11 in many aspects, including data corruption ratio and power efficiency. Copyright © 2008 John Wiley & Sons, Ltd.

Abstract It is important to provide quality of service (QoS) guarantees if we want to support multimedia applications over wireless networks. In this paper, considering the features of tiering in sectored cellular networks, we propose a novel scheme for bandwidth reservation to approach QoS provisioning. By predicting the movement of each connection, the reserving of bandwidth is only required in needful neighboring cells instead of in all neighboring cells. In addition, an admission control mechanism incorporated with bandwidth borrowing assists in distributing scarce wireless bandwidth in more adaptive way. Through mathematical analysis, we proof the advantages of tier‐based approach and the bound for the selection of tiered boundary. The simulation results also verify that our scheme can achieve superior performance than traditional schemes regarding no bandwidth reserving, fixed bandwidth reserving, and bandwidth borrowing in sectored cellular networks when performance metrics are measured in terms of the connection dropping probability (CDP), connection blocking probability (CBP), and bandwidth utilization (BU). Copyright © 2008 John Wiley & Sons, Ltd.

A new network transmission framework highlighting the concept of cross-layer co-design for packet loss reduction is presented in this article. The proposed framework takes into account the correlation between the packet loss rate and the packet size under a given channel condition, and assigns different tasks to layers involved: physical/ MAC layer for monitoring current channel statuses, IP layer for fragmentation, transport layer for bandwidth allocation, and application for the selection of coding method and compression rate. Our proposed scheme can detect current channel conditions and dynamically adjusts parameters in each layer to match the changing channel status. Especially, when a mobile node is conducting a handover process,the radio signals and channel conditions will change rapidly. The proposed scheme can accommodate the variation quickly and prepares adequate data for corresponding node in advance, so that we can reduce the handover delay and make seamless handover possible.

In infrastructure wireless networks, the wireless hop can be considered as another hop of the transmission path. With the rapid growth of wireless traffics, the future wireless network is expected to provide services for heterogeneous data traffics with different quality of service (QoS) requirements. Most proposed schemes do not have adaptive mechanisms to deal with the environment changes. In real situation, bandwidths, error rates and loss rates of wireless links vary frequently. We will base on the differentiated service model and propose a wireless differentiation (WD) scheme for user datagram protocol (UDP) flows and a wireless differentiation with prioritised ACK scheme for connections with transmission control protocol (TCP) flows. Both schemes provide QoS support for IEEE 802.11b and do not change the basic access mechanism of IEEE 802.11b.

In wireless/mobile networks, users freely change their service points, while they are communicating with other users. In order to support the mobility of mobile nodes, Mobile IPv6 (MIPv6) is proposed by the Internet Engineering Task Force (IETF), in which a mobile node must inform its home agent the binding of its home address and the current care-of-address (CoA). The home agent forwards packets to CoA when it receives packets for the mobile node. There is a significant problem in MIPv6 due to its inability to support micromobility caused by long delays and high packet losses during a handoff. Hierarchical Mobile IPv6 (HMIPv6) is proposed to separate mobility into micromobility [within one domain or within the same mobility anchor point (MAP)] and macromobility (between domains or between MAPs). HMIPv6 reduces handoff latency by employing a hierarchical network structure and minimizing the location update signaling with external network. The two-layer network structure of HMIPv6 is very suitable for supporting the vertical handoff and the horizontal handoff in wireless overlay networks. Wireless overlay networks can consist of two layers. Functions of access routers can be implemented in low-layer networks and functions of MAPs can be implemented in high-layer networks. Or we can have an access router collocated with a low-layer network and an MAP collocated with a high-layer network. In either way, the micromobility in HMIPv6 is equal to the horizontal handoff in wireless overlay networks and the macromobility in HMIPv6 is equal to the vertical handoff in wireless overlay networks. However, a significant delay still occurs in macromobility management. This paper considers handover operations. We analyze the handoff delay and find that the DAD time represents a large portion of handoff delay. We also note that we can assign a unique on-link care-of address (LCoA) to each mobile node and switch between one-layer IPv6 and two-layer IPv6 addressing. By this way, we propose a Stealth-time HMIP (SHMIP) which can reduce the effect of the DAD time on the handoff delay and, thus, reduce the handoff time significantly. To further reduce packet losses, we also adopt prehandoff notification to request previous MAP to buffer packets for the mobile node. By simulations, we show that the proposed scheme can realize low handoff delays and low packet losses during macromobility.

Many mechanisms based on bandwidth reservation have been proposed in the literature to decrease connection blocking probability and connection dropping probability in cellular communications. The handoff events occur at a much higher rate in sectored cellular networks than in traditional cellular systems. An efficient bandwidth reservation mechanism for the neighboring cells is therefore critical in the process of handoff during the connection of multimedia calls to avoid the unwillingly forced termination and waste of limited bandwidth in the sectored cellular communications, particularly when the handoff traffic is heavy. A self-adaptive bandwidth reservation scheme which employs a neural fuzzy bandwidth-reserving estimator, is proposed, and the simulation results show that our scheme can achieve superior performance than traditional fixed bandwidth-reserving scheme in sectored cellular networks when performance metrics are measured in terms of the new call blocking probability and the forced termination probability.

We investigate packet discarding schemes for transport control protocol (TCP) over asynchronous transfer mode with guaranteed frame rate service. In this letter, we propose the selective weighted fair allocation (SWFA) scheme, which discards packets from selected sessions. 15 TCP connections with equal minimum cell rate (MCR) and unequal MCRs are simulated. The SWFA scheme with per-virtual connection (VC) queuing is compared with early packet discard (EPD) with first-in, first-out queuing, EPD with per-VC queuing, and differential fair buffer allocation with per-VC queuing. Our results show that SWFA with per-VC queuing achieves significant enhancement on throughputs, goodputs, and fairness.

UBR connections of ATM networks provide no service guarantees. Schemes such as EPD (early packet discard) or PPD (partial packet discard), though improving the goodputs of packets over UBR (unspecified bit rate), cannot guarantee a MCR (minimum cell rate) for each flow. Guerin and Heinanen (1996) proposed a new service, which was originally called "UBR+" and is now called GFR (guaranteed frame rate), to assure a MCR for each flow. We investigate packet-discarding schemes for TCP over ATM with GFR service. We propose the SWFA (selective weighted fair allocation) scheme, which discards packets from selected sessions. 15 TCP connections with equal MCR and unequal MCRs are simulated. The SWFA scheme with per-VC queuing is compared with EPD with FIFO queuing, EPD with per-VC queuing and DFBA (differential fair buffer allocation) with per-VC queuing. Our results show that SWFA with per-VC queuing achieves significant enhancement on throughputs, goodputs, and fairness.

The Internet is increasingly used to deliver multimedia services. Since there are heterogeneous receivers and changing network conditions, it has been proposed to use adaptive rate control techniques such as layered video multicast to adjust the video traffic according to the available Internet resources. A problem of layered video multicast is unable to provide fair bandwidth sharing between competing video sessions. We propose two schemes, layered video multicast with congestion sensitivity and adaptive join-timer (LVMCA) and layered video multicast with priority dropping (LVMPD), to achieve inter-session fairness for layered video multicast. Receiver driven layered multicast (RLM), layer-based congestion sensitivity, LVMCA, and LVMPD are simulated and compared. The results show that both schemes, especially LVMPD, are fairer and have shorter convergence time than other two schemes.

In this paper, we present an algorithm for performing permutations of messages on multistage interconnection networks. Permutations of messages are needed in many parallel algorithms. The proposed algorithm is feasible for any networks that can connect each input to each output using a set of N nonblocking connections, where N is the number of ports on the network. Messages are segmented into N submessages that are sent independently in each time step. For any permutation, the settings of switches are changed with fixed patterns. Partitioning of the network into independent subnetworks is also supported, each capable of simultaneously routing a different permutation.

Multicast delivery is important in multimedia applications. These data can be heavily time-critical in terms of maximum delay and jitter. To satisfy these applications, the network needs to have an efficient multicasting mechanism using the true capability of ATM networks. In the original solution, a separate connection can be set up from the source to each group node, and therefore it needs O(N/sup 2/) connections, where N is the number of nodes in a group. Instead, we propose a bi-directional multipoint-to-multipoint multicast scheme, a synchronizer-dispatcher channel-based multicast with round-robin access (SDRAM), for ATM networks, which uses a single tree for a multicast group consisting of multiple participants that are either senders or receivers or a mix of both. Simulation results show the mean delays and the mean inter-PDU delays of SDRAM are not sensitive to mean PDU size while the resequencer is very sensitive to that one.