Vehicle Edge Computing (VEC) has emerged as an efficacious paradigm that supports real-time, computation-intensive vehicular applications. However, due to the highly dynamic nature of computing node topology, existing scheduling algorithms need to more effectively apprehend the characteristics of fine-grained task topologies and network topologies. Moreover, they require significant communication overhead and training costs, making them inadequate for fine-grained task scheduling in vehicular networks. In response, our research explores fine-grained task scheduling issues within VEC scenarios, proposing a scheduling algorithm based on Graph Neural Networks and Federated Learning (FL-GNN). This algorithm maintains a global scheduling model that periodically aggregates local scheduling models deployed on Roadside Units (RSUs) and high-performance vehicles. Furthermore, to enhance the model’s ability to perceive topology and expedite the convergence rate, we incorporate a graph neural network layer in each local model to preprocess the raw state of the VEC environment. Lastly, we construct a simulation platform and implement multiple competitive solutions, demonstrating the superiority of the FL-GNN algorithm in aspects such as reducing the average task delay, balancing the load, and improving the task scheduling success rate.

精准地预判网络流量变化趋势可以帮助运营商准确预估网络的使用情况,合理分配并高效利用网络资源,以满足日益增长且多样化的用户需求.以深度学习算法在网络流量预测领域的进展为线索,阐述了网络流量预测的评价指标和目前公开的网络流量数据集及应用,具体分析了网络流量预测中常用的深度信念网络、卷积神经网络、循环神经网络和长短时记忆网络共四种深度学习方法,并重点介绍了近年来针对不同问题所提出的改进神经网络模型,总结了各模型特点及应用场景.最后对网络流量预测未来发展进行了展望.

In a cloud distributed system, machine failure or region failure is a very common scenario. Data replication is a key technique for ensuring data availability. However, Objects are usually assumed independently by distributed systems, despite, a user-level task typically requests multiple data objects. This paper studies the effect of data placement on the availability of user-level tasks from a theoretical perspective, and finds the best and the worst placements which can provide the highest and the lowest availability for user-level tasks in a cloud distributed system. This paper also gives a novel algorithm called SPOverlap (S Parts Overlap), which provides a tradeoff between task availability and other system performance.

The Market of Data is an environment where data are reasonably deal with. Some data in the market of data are large and hard to analyze. How to efficiently analyze and organize such large scale data in the market of data is a difficult problem. When using Hadoop to analyze these massive data, if input data of a data mining task are not locally available in a processing node, data have to be migrated via network interconnects to node that performs the data processing operations. These data movement obviously has a bad effect on system performance. In this paper, we propose BRPS (Big data Replicas Placement Strategy), a strategy that improves data intensive tasks parallel execution performance by reducing data movement across multiple machines. The simulation results show that BRPS can greatly reduce the data movement cost and promote workload balance slightly.

With the increasing number of location based services (LBS) and GPS-enabled devices, there are more and more Web based GPS trajectory applications. Since the huge volume of spatiotemporal trajectory data brings about heavy burden for data storing and querying, it is of great importance to organize the trajectory data efficiently to satisfy the indexing demand of many LBS applications. Traditional methods in spatiotemporal database always incur a data search in a large range while users query about trajectories in a particular spatiotemporal region in Web based applications. In this paper, we develop a data frame based spatiotemporal indexing algorithm for moving objects, which is aimed to achieve high efficiency in spatiotemporal retrieval for trajectories. The method first allocates points from different objects of the same timestamp into the identical data frame to construct temporal index. Then it partitions the frames into groups, each group starts with a static frame followed by several dynamic frames. Finally, the method divides the space into blocks based on Geohash algorithm. In this way, temporal and spatial index are combined to achieve high efficiency during the trajectory retrieval. Experiments on real large scale trajectory datasets demonstrate that the proposed algorithm has superiority over traditional methods in speeding up the spatiotemporal retrieval and is quite qualified to the characteristics of Web based LBS applications.

In view of the complexity of the network and terminal in heterogeneous environment,and the mutability of playback progress in the video-on-demand service that is caused by user random seeking, an Scalable Video Coding ( SVC) fragment schedule algorithm based on user behavior characteristic is proposed. In the proposed algorithm, two types of windows are designed. One is playback window based on current playtime to ensure order data continues, the other one is anchor window designed with data prefetching,which is based on user random seeking following the Weibull distribution. The Layer-by-Layer( LL) schedule strategy is utilized in playback window and the first anchor window to ensure the timeliness of data, and the rarest-first strategy is used in the other anchor windows to balance the fragment distribution of the whole system. Simulation results in OverSim show that,compared with current LL schedule algorithm and weighted schedule algorithm,the proposed algorithm can improve the fragment scheduling performance,shortens the response time delay,reduces the server load,and improves the quality and fluency of the user in watching video.

In this paper, we propose a data mining based publish/subscribesystem (DMPSS). First, the data mining technology is used to find attributes that are usually subscribed together, e.g. frequent itemset. Then subscriptions and events are installed by frequent itemsets contained in them. If subscriptions and events don't contain any frequent itemset, they are delivered to specified RPs (rendezvous points) for matching. The usage of frequent itemsets provides two advantages to DMPSS. First, it achieves even matching load distribution on RPs. Second, it reduces the event publication cost. The performance of DMPSS is evaluated by simulations. The experimental results show that DMPSS realizes even matching load distribution, and it reduces the overhead for message transmission and latency dramatically.

With the increasing number of GPS-enabled devices, the huge volume of spatiotemporal trajectory data brings about heavy burden for data storing, transmitting and processing. Compressing large scale trajectories is in urgent need. In this paper, we develop an algorithm for trajectory compression based on non-uniform quantization (TCNQ). It is aimed to achieve high compression ratio of large scale trajectory data when lacking geographic context. The method first converts spatiotemporal trajectories into strings by encoding differential coordinates with non-uniform quantization. Then run-length coding is performed to remove redundant points from the trajectories. In this way, quantization and trajectory simplification are combined in the encoding procedures to achieve high compression ratio without losing too much information. Experiments on real large scale trajectory datasets demonstrate that the proposed algorithm has superiority over traditional methods in compression ratio as well as deviation control.

We propose a playback length changeable 3D video data chunk segmentation algorithm.A novel hybrid-priority based 3D video P2P data scheduling algorithm is presented.Bandwidth-adaptive 3D video P2P streaming for heterogeneous networks is implemented.The proposed method obtains superior error-resilience and network utilization performance. 3D video distribution over P2P networks has been thought as a promising way for 3D video entering home. The convergence of scalable 3D video coding and P2P streaming can provide diverse 3D experiences for heterogeneous clients with high distribution efficiencies. However, the conventional chunk segmentation and scheduling algorithms originally aiming at the non-scalable 2D video streaming are not very efficient for scalable 3D video streaming over P2P networks due to the particular data characteristics of scalable 3D video. Based on this motivation, this paper first presents a playback length changeable 3D video chunk segmentation (PLC3DCS) algorithm to provide different error resilience strengths to video and depth as well as layers with different importance levels in the 3D video transmission. Then, a hybrid-priority based chunk scheduling (HPS) algorithm is proposed to be tied in with the proposed chunk segmentation algorithm to further promote the overall 3D video P2P streaming performance. The simulation results show that the proposed PLC3DCS algorithm with the corresponding HPS can increase the success delivery rates of chunks with more important levels, and further improve the user's quality of 3D experience.

In this paper, we propose a QoE Space (QS) based QoE Adaptation (QSQA) algorithm for SVC based peer-to-peer (SVC-P2P) video streaming systems. The QS records the users' experiences in terms of Mean Opinion Scores (MOS) with different content type, packet loss rate, spatial resolution, frame rate and video compression Quantization Parameter (QP). The two features of QS, the video content classification and the linear interpolation of MOS, make it can be universally used for QoE prediction in video streaming systems. The prediction accuracy of QS is also validated from these two aspects. By utilizing the initial QoE adaptation and progressive QoE adaptation, the QSQA algorithm determines layer subscription for client according to network environments, terminal conditions and video content characteristics. Extensive subjective experimental results of SVC-P2P video streaming show that QSQA obviously outperforms quality adaptation algorithms that based on Layer-by-Layer and JSVM bit stream extraction.

Three key performance indexes of distributed publish/subscribe systems are latency, load balance and flow cost. In this paper, we propose Adaptable Publish/Subscribe Architecture (APSA) for content-based publish/subscribe system, and it is built on structured P2P networks. In APSA, We use subscriptions covering, merging and binding, and two-layer delivering tree strategies to balance the load and reduce latency, so as to get a good performance. The experimental results show that APSA can provide modifiable load balance with relative low latency and low network flow cost which fit for publish/subscribe systems.

Scalable 3D video P2P streaming systems can supply diverse 3D experiences for heterogeneous clients with high efficiencies. Data characteristics of the scalable 3D video make the P2P streaming efficiency more depends on the data segmentation algorithm. However, traditional data segmentation algorithm is not very appropriate for scalable 3D video P2P streaming systems. In this paper, we propose a Playback Length Changeable 3D video Segmentation (PLC3DS) algorithm. It considers the particular source-data characteristics of scalable 3D video, and provides different error resilience strengths to video and depth as well as layers with different importance levels in the transmission. The simulation results show that the proposed PLC3DS algorithm can increase the success delivery rates of chunks in more important layers, and further improve the 3D experiences of the client. Moreover, it improves the network utilization ratio remarkably.

This paper classified SVC-based P2P steaming systems into three categories according to the architecture of data transmission,including tree-based,hybrid,and mesh-based.Then it thoroughly investigated the advances in the above mentioned three types systems,especially the mesh-based SVC-based P2P streaming systems,and discussed deficiencies in the existing works and issues to be further studied in the future.

The equal playback length segmentation algorithm, widely used in P2P streaming systems based on scalable video coding (SVC), was studied to find an approach for P2P streaming systems to transmit and share video data in a helerogeneous environment. The results show that this algorithm does not consider the data characteristics of SVC adequately, and it may decrease video qualities on clients seriously when packet loss rate is high. To deal with this problem, the study proposed an unequal playback length segmentation algorithm. It adjusts the transmission success possibility of layers by changing the playback length of chunks, and guarantees the successful transmission of layers with high importance levels firstly. The proposed algorithm was evaluated by experiments and simulations and the results show that, compared with the equal playback length segmentation algorithm, the new proposed algorithm will smooth the transmission time of chunks, and increase the transmission success possibility of layers with high importance levels. Moreover, it performances better in different network environments and it is suitable to heterogeneous networks.

The PLCS (Playback Length Changeable Segmentation) algorithm was proposed to improve the success delivery rate of chunks in some important layers including the base layer for SVC based P2P streaming systems. In this paper, we make further research and propose a corresponding chunk scheduling algorithm named PLCCS (Playback Length Changeable Chunk Scheduling). In PLCCS, there are two scheduling windows on clients, including General Window and Emergent Window. Different scheduling strategies are adopted in the two windows to ensure the efficiency of data distribution as well as the playback fluency. Moreover, an adaptive strategy is used in General Window to accommodate the change of the network state. We evaluate the performance of PLCCS through simulations. The results show that compared with two representative EPLS based chunk scheduling algorithms, PLCCS can achieve better video quality on clients by at least 40%, and it decreases the useless packet ratio on clients by at least 15 times. Moreover, the adaptive strategy used in PLCCS makes it suitable to be applied in SVC based P2P streaming systems in heterogeneous network environments.

We propose a novel segmentation algorithm TSESS and a novel chunk scheduling algorithm TSESCS for SVC based P2P streaming systems. TSESS splits each layer into equal sized chunks and tags each chunk with a time-stamp. TSESCS is based on TSESS and adopts two scheduling windows with adaptive strategies. Compared with traditional segmentation algorithm and chunk scheduling algorithms, our TSESS and TSESCS algorithms can achieve better video quality on clients in heterogeneous network environments.

Due to the architecture and characteristics of 3G cellular networks, peer selection algorithms with the idea of traffic localization are not suitable for mobile P2P systems in 3G cellular networks, and load balance on the cells should be taken into account. In this paper, we propose a peer selection algorithm named CFLB for mobile P2P systems in 3G cellular networks. The CFLB algorithm always chooses a peer with the highest available uplink bandwidth from a cell with the lowest traffic load until the number of peers is reached. Simulation results indicate that the CFLB algorithm can achieve excellent load balance on the cells in 3G cellular networks while ensuring favorable peer performance.