Search

To solve the problem of huge pressure on network operation caused by the surge of 5G mobile data traffic and the continuous emergence of new network applications, a traffic scheduling system of the 5G core network user plane was designed and implemented, including network state information perception subsystem and routing decision subsystem.In the network state information perception subsystem, traditional in-band network telemetry methods incurred high bandwidth overhead and were not specifically designed for wireless network systems.Furthermore, the application of in-band network telemetry in the 5G core network faced challenges such as low measurement accuracy and the inability to guarantee QoS.A 5G core network user plane state information perception scheme based on in-band network telemetry was proposed.The telemetry information was inserted into the extended header of GTP-U message realize the measurement of network state along the path.The routing decision subsystem implemented the traffic scheduling algorithm based on the improved ant colony algorithm.The update method of the pheromone function was promoted by the perceived status information to complete the routing decision based on the real-time network state.The deployment test results showed that the network state information perception subsystem can perceive network information normally, and the routing decisions were superior to traditional routing algorithms in terms of delay, throughput and packet loss rate.

Federated learning (FL) allows a number of parties collectively train models without revealing private datasets. There is a possibility of extracting personal or confidential data from the shared models even-though sharing of raw data is prevented by federated learning. Secure Multi Party Computation (MPC) is leveraged to aggregate the locally-trained models in a privacy preserving manner. However, it results in high communication cost and poor scalability in a decentralized environment. We design a novel communication-efficient MPC enabled federated learning called CE-Fed. In particular, the proposed CE-Fed is a hierarchical mechanism which forms model aggregation committee with a small number of members and aggregates the global model only among committee members, instead of all participants. We develop a prototype and demonstrate the effectiveness of our mechanism with different datasets. Our proposed CE-Fed achieves high accuracy, communication efficiency and scalability without compromising privacy.

Abstract Vehicular edge computing (VEC) is a promising paradigm to offload resource-intensive tasks at the network edge. Owing to time-sensitive and computation-intensive vehicular applications and high mobility scenarios, cost-efficient task offloading in the vehicular environment is still a challenging problem. In this paper, we study the partial task offloading problem in vehicular edge computing in an urban scenario. Where the vehicle computes some part of a task locally, and offload the remaining task to a nearby vehicle and to VEC server subject to the maximum tolerable delay and vehicle’s stay time. To make it cost-efficient, including the cost of the required communication and computing resources, we consider to fully exploit the vehicular available resources. We estimate the transmission rates for the vehicle to vehicle and vehicle to infrastructure communication based on practical assumptions. Moreover, we present a mobility-aware partial task offloading algorithm, taking into account the task allocation ratio among the three parts given by the communication environment conditions. Simulation results validate the efficient performance of the proposed scheme that not only enhances the exploitation of vehicular computation resources but also minimizes the overall system cost in comparison to baseline schemes.

With proliferation of smart devices and wireless applications, the recent few years have witnessed data surge. These massive data needs to be stored, transmitted, and processed in time to exploit their value for decision making. Conventional cloud computing requires transmission of massive amount of data in and out of core network, which can lead to longer service latency and potential traffic congestion. As a new platform, mobile edge computing (MEC) moves computation and storage resources to edge network in proximity to the data source. With MEC, data can be processed locally, and thus mitigate issues of latency and congestion. However, it is very challenging to reap the benefits of MEC everywhere due to geographic constraints, expensive deployment cost, and immoveable base stations. Because of easy deployment and high mobility of unmanned aerial vehicles (UAVs), air-ground integrated mobile edge networks (AGMEN) is proposed, where UAVs are employed to assist the MEC network. Such an AGMEN expects to provide MEC services ubiquitously and reliably. In this article, we first introduce the characteristics and components of UAV. Then, we will review the applications, key challenges, and current research technologies of AGMEN, from perspectives of communication, computation, and caching, respectively. Finally, we will discuss some essential research directions for AGMEN.

The number of mobile applications(APPs) has increased dramatically with the development of mobile Internet. It becomes challenging for users to identify these APPs they are really interested in. Existing mobile APP recommendation methods focus on learning users' preference and recommending high visibility APPs. However, some low visibility APPs may satisfy users and even surprise them. If those low visibility APPs have the opportunity to show to the user, they will not only improve the user's satisfaction, but also provide a fair competitive market for APP providers. Furthermore, it will improve the vitality of the APP market. To this end, we present a fairness-aware APP recommendation method named FARM. The principal study of this method emphasizes on the fairness issue during the recommendation process. In this method, APP candidates are divided into high visibility and low visibility APPs, and implement recommendation algorithm respectively. For low visibility APPs, we set a fairness factor for everyone, and use the user's latest feedback to make a dynamic adjustment. Based on the fairness factor, the recommendation is implemented by roulette-wheel. For high visibility APPs, we employ the fuzzy analytic hierarchy process to implement the recommendation. The evaluation results show that FARM outperforms baselines in terms of recommendation fairness.

Abstract Content delivery network (CDN) has gained increasing popularity in recent years for facilitating content delivery. Most existing CDN-based works upload the content generated by mobile users to the cloud data center firstly. Then, the cloud data center delivers the content to the proxy server. Finally, the mobile users request the required content from the proxy server. However, uploading all the collected content to the cloud data center increases the pressure on the core network. In addition, it also wastes a lot of bandwidth resources because most of the content does not have to be uploaded. To make up for the shortcomings of existing CDN-based works, this article proposes an edge content delivery and update (ECDU) framework based on mobile edge computing architecture. In the ECDU framework, we deploy a number of content servers to store raw content collected from mobile users, and cache pools to store content that frequently requested at the edge of the network. Thus, it is not necessary to upload all content collected by mobile users to the cloud data center, thereby alleviating the pressure of the core network. Based on content popularity and cache pool ranking, we also propose edge content delivery (ECD) and edge content update (ECU) schemes. The ECD scheme is to deliver content from cloud data center to cache pool, and the ECU scheme is to mitigate the content to appropriate cache pools in terms of its request frequency and cache pool ranking. Finally, a representative case study is provided and several open research issues are discussed.

Dynamic adaptive video streaming over HTTP (DASH) is widely studied and has been adopted in modern video players to ensure user quality of experience (QoE). In DASH, adaptive bitrate control is a key part whose ultimate goal is to maximize video bitrate while minimizing rebuffering. Throughput prediction plays an important role in helping select the proper video bitrate dynamically. In this paper, we studied the influence of throughput prediction on adaptive video streaming. Because the real-world network is dynamic, different methods need to be tested with large-scale deployments and analyzed statistically. However, this is difficult in academic research. Therefore, we established a reproducible trace-based emulation environment, which enables us to compare different methods quantitatively under the artificially same condition, with limited experiments. The throughput prediction methods are implemented into DASH to evaluate the effect on QoE for video streaming. The results indicate that the prediction method using long short-term memory (LSTM) performs better than the other methods. However, throughput prediction alone is not enough to ensure high QoE. To further improve the QoE, we proposed the decision map method (DMM), where the buffer occupancy is also incorporated to make a selection. By using this decision map, the choice of bitrate can be smarter than that when only prediction information is used. The total QoE is further improved by 32.1% in the ferry trace, which shows the effectiveness of DMM in further improving the performance of throughput prediction in adaptive bitrate control.

With the increasing number of in-vehicle sensors and the adoption of various communication interfaces, vehicles have become an ambient sensing platform for many applications such as traffic and weather monitoring, urban surveillance, road safety, and driver behavior analysis. Furthermore, a large number of vehicles can be coordinated to sense large-scale information precisely and efficiently. Such coordination requires enormous communication, computing, and power resources. In this paper, we propose a virtual vehicle (VV) coordination approach based on the theory of group consensus to sense environment information efficiently. Specifically, we first propose a discovery algorithm to find the optimal VV groups. Then, based on multi-group consensus theory, we devise a coordination algorithm to control VV groups to achieve multi-group coordination by adjusting the communication relationships among VVs. Finally, the precise environment information is acquired through the coordination of multiple VV groups. Extensive simulations are provided to demonstrate the effectiveness of the proposed schemes.

Group recommendation generates a ranked list of recommendations for a group of users. Point-of-interests (POIs) group recommendation aims to suggest the most agreeable meeting places for a group of users. Although there are a lot of studies on group recommendation for POIs, few studies take into account the rationality of location for the whole group. In this paper, we propose a novel POI group recommendation method which factors into the rationality of the location and the intra-group influence when making group decisions. We take into account the importance of location in POI recommendations and employ distance-based pre-filtering and distance-based ranking adjustment to improve recommendation satisfaction. We have conducted extensive experimental evaluations of the proposed method via a realworld data set, which is prepared from 1 375 024 Beijing POI comment records hosted by a review website. Comprehensive experimental results show that our proposed POI group recommendation method outperforms other representative ones in terms of global satisfaction and distance satisfaction, even in the context of individual recommendation.

Mobile edge computing (MEC) provides a promising approach to significantly reduce network operational cost and improve quality of service (QoS) of mobile users by pushing computation resources to the network edges, and enables a scalable Internet of Things (IoT) architecture for time-sensitive applications (e-healthcare, real-time monitoring, and so on.). However, the mobility of mobile users and the limited coverage of edge servers can result in significant network performance degradation, dramatic drop in QoS, and even interruption of ongoing edge services; therefore, it is difficult to ensure service continuity. Service migration has great potential to address the issues, which decides when or where these services are migrated following user mobility and the changes of demand. In this paper, two conceptions similar to service migration, i.e., live migration for data centers and handover in cellular networks, are first discussed. Next, the cutting-edge research efforts on service migration in MEC are reviewed, and a devisal of taxonomy based on various research directions for efficient service migration is presented. Subsequently, a summary of three technologies for hosting services on edge servers, i.e., virtual machine, container, and agent, is provided. At last, open research challenges in service migration are identified and discussed.

With the development of converged all-IP networks, new network services-supporting, among others, image, audio, and video-with high diversity emerge, which brings huge pressure and fierce competition to service providers because of increasing choices for users. Assessing end user's quality of experience (QoE) has become one of the main issues to meet high user requirements and satisfaction under any circumstances, which means receiving any service, anytime, anywhere, on any device, by any media and networking technology, across any operator domain. In this paper, we provide an overview of the field of QoE assessment approaches including three aspects: subjective assessment, objective assessment, and hybrid assessment. The discussion includes the definition, influence factors, and quantification approaches for QoE. Then, these three assessment methodologies are surveyed and compared across a number of factors. This paper also introduces challenges and opportunities in the area of QoE assessment.

In this paper, the energy-efficient massive MIMO design problem is investigated in both single-cell and multi-cell scenarios with the aim of selecting appropriate parameters to maximize the overall system energy utilization. To this end, a more realistic parameter selection model is established by considering the channel estimation error and pilot contamination (PC). In the proposed model, we not only optimize the number of antennas, the number of users, and the transmission power to balance the improved radiated energy efficiency and the increased circuit power consumption, but also optimize the length of pilot sequences to make a desired trade-off between improving channel estimation accuracy and reserving more resources for data transmission. We first focus on the single-cell scenario, and strictly prove the quasi-concavity of the objective function in each dimension after suitable variable transformation, based on which an alternative optimization plus bisection searching (AO-BS) algorithm is proposed to solve the problem quickly. Then, the research is extended into the multi-cell scenario by increasingly considering PC and multi-cell interference. Since PC is related to the relative pilot length, i.e., the ratio between the length of pilot sequences and the number of users, the formulated problem is difficult to deal with directly. For effective solving, we divide the original problem into multiple subproblems according to the setting of pilot reusing factor, and employ the proposed AO-BS algorithm to solve each subproblem. Through numerical simulations, it is observed that the proposed AO-BS algorithm converges rapidly and is almost able to achieve the globally optimal solution. Meanwhile, from the perspective of energy efficiency maximization, massive MIMO is more preferable in the case of large cell coverage, and selecting a large relative pilot length is beneficial, always resulting in higher energy utilization in both single-cell and multi-cell scenarios.

The Internet of Things (IoT) generates lots of high-dimensional sensor intelligent data. The processing of high-dimensional data (e.g., data visualization and data classification) is very difficult, so it requires excellent subspace learning algorithms to learn a latent subspace to preserve the intrinsic structure of the high-dimensional data, and abandon the least useful information in the subsequent processing. In this context, many subspace learning algorithms have been presented. However, in the process of transforming the high-dimensional data into the low-dimensional space, the huge difference between the sum of inter-class distance and the sum of intra-class distance for distinct data may cause a bias problem. That means that the impact of intra-class distance is overwhelmed. To address this problem, we propose a novel algorithm called Hierarchical Discriminant Analysis (HDA). It minimizes the sum of intra-class distance first, and then maximizes the sum of inter-class distance. This proposed method balances the bias from the inter-class and that from the intra-class to achieve better performance. Extensive experiments are conducted on several benchmark face datasets. The results reveal that HDA obtains better performance than other dimensionality reduction algorithms.