Microservice architecture is gradually replacing traditional monolithic applications, but its modular design introduces new security challenges in service interactions. To ensure the security of the entire microservices system, it is crucial to implement strict network policies that prevent a compromised service from maliciously exploiting or affecting other properly functioning microservices. However, the dynamic and large-scale nature of microservices makes traditional manual policy configuration impractical. This paper presents a network policy generation method based on service topology. Our approach comprises two fundamental techniques: (1) a real-time service topology generation method that leverages configuration files and network plugins to dynamically capture service interactions, and (2) an automatic network policy generation method based on the derived service topology. The effectiveness of the proposed method is validated through experiments on three popular microservice applications. The results demonstrate that our method can generate fine-grained and efficient network policies, reducing the time overhead by an average of 38

With the development of Artificial Intelligence for IT Operations(AIOps), an increasing number of log analysis approaches have been proposed to guarantee the stability and reliability of large-scale systems. However, traditional approaches still exist some limitations. First, they cannot adapt to the changing log statements. Second, most of them are based on large-scale labeled data. Third, most of their frames are fixed. To overcome these limitations, we propose OpenLog based on meta-learning and few-shot learning. Experimental results show that OpenLog can achieve an accuracy of 95.8% on detecting unknown anomaly classes, and an accuracy of 93.3% in anomaly classification on unbalanced logs with unknown anomaly classes. In the few-shot classification, OpenLog can still achieve an accuracy over 80%.

With the development of Artificial Intelligence (AI), Internet of Things (IoT), cloud computing, new-generation mobile communication, etc., digital transformation is changing the technical architecture of IT systems. It brings more requirements for performance and reliability. The traditional human-dependent development and maintenance methods are overwhelmed, and need to transform to Artificial Intelligence for IT Operations (AIOps). As one of the most useful data resources in IT system, the log plays an important role in AIOps. There are many research on enhancing log quality, analyzing log structure, understanding system behavior, helping users to mine the effective information in logs. Based on the characteristics of logs and different strategies, this paper reviews and categorizes the existing works around the three key processes in the log processing framework of log enhancement, log parsing, and log analysis in academia, and establishes evaluation indicators for comparison and summary. Finally, we discussed the potential directions and future development trends.

Increased awareness of privacy protection has led to a surge in the volume of encrypted traffic, which creates a heavy burden for efficient network management (e.g. quality-of-service guarantees). The opacity of encrypted traffic essentially requires high computational overheads to make traffic classification, which is even worse when encrypted traffic surges. However, existing deep learning approaches sacrifice the efficiency to obtain high-precision classification results, which are no longer suitable for scenarios with large volumes of encrypted traffic. In this paper, a lightweight and online approach implemented as MATEC is proposed. The way we optimize the classification process follows the “Maximizing the reuse of thin modules” design principle. The multi-head attention and the convolutional network are adopted in the thin module. Attributed to the one-step interaction of all packets and the parallel computing of the multi-head attention mechanism, a key advantage of our model is that the number of parameters and running time are significantly reduced. In addition, the effectiveness and efficiency of convolutional networks have been proved in traffic classification. Comparisons to the existing state-of-the-art models on three typical datasets demonstrate that the proposed MATEC model has higher accuracy and running efficiency. In addition, the number of parameters is reduced to 1.8% of the state-of-the-art models and the training time halves.

Logs are used to record runtime states and significant events for a software system. They are widely used for anomaly detection. Logs produced by most of the real-world systems show clear characteristics of imbalanced data because the number of samples in different classes varies sharply. The distribution of imbalanced data makes the anomaly classifier bias toward the majority class, so it is difficult for a classifier to learn to detect anomalies correctly. Most existing methods for log-based anomaly detection ignore this important problem, so they perform poorly on real-world systems. In this paper, we propose a context-aware method named AllContext for anomaly detection with imbalanced log data. AllContext transforms each log event into a vector, which contains not only the semantic information of each word, but also the semantics of the region where each word is located. Such rich semantic information enables our method to understand the imbalanced log data better and deeper. We conduct extensive experiments on multi-class and binary imbalanced log datasets. The accuracy of the proposed AllContext is more than twofold of that by a baseline state-of-the-art. To evaluate the robustness of the proposed method, we assess AllContext on the imbalanced unseen log data, where all samples in the test dataset do not exist in the training dataset, and the accuracy achieved by AllContext reaches 0.98. The experiments show that the proposed solution achieves accurate results on both imbalanced and unseen log data.

The fast growth of encrypted traffic puts forward burning requirements on the efficiency of traffic classification. Although deep learning models perform well in the classification, they sacrifice the efficiency to obtain high-precision results. To reduce the resource and time consumption, a novel and lightweight model is proposed in this paper. Our design principle is to “maximize the reuse of thin modules A thin module adopts the multi-head attention and the 1D convolutional network. Attributed to the one-step interaction of all packets and the parallelized computation of the multi-head attention mechanism, a key advantage of our model is that the number of parameters and running time are significantly reduced. In addition, the effectiveness and efficiency of 1D convolutional networks are proved in traffic classification. Besides, the proposed model can work well in a real time manner, since only three consecutive packets of a flow are needed. To improve the stability of the model, the designed network is trained with the aid of ResNet, layer normalization and learning rate warm up. The proposed model outperforms the state-of-the-art works based on deep learning on two public datasets. The results show that our model has higher accuracy and running efficiency, while the number of parameters used is 1.8% of the 1D convolutional network and the training time halves.

•A novel deep learning method for efficient HTTPS traffic classification.•Bidirectional GRU to extract forward and backward features of byte sequences.•Attention mechanism to focus on useful features for traffic classification.•Transfer learning is adopted for re-training the model quickly.•Reduce the number of parameters of the model by 7 times compared with the baseline.

The IP address attribution information includes the geographical information, the network routing information, the agency information, Internet Content Provider (ICP) information, etc. Nowadays, the attribution information is important to the network traffic engineering, which needs to be obtained in real time in network traffic analysis system. The existing proposed methods for IP address attribution information lookup cannot be employed in actual systems efficiently due to their low scalability or bad performance. They cannot address the backbone network's requirements for real-time IP address attribution information lookup, and most lookup methods do not support custom IP address attribution lookup. In response to these challenges, we propose a novel high-speed approach for IP address attribution information lookup. We first devise a data structure of IP address attribution information search tree (PATIP-TREE) to store custom IP address attribution information. Based on the PATIP-TREE, an effective algorithm for IP information lookup is proposed, which can support custom IP addresses attribution information lookup in real time. The experimental results show that our method outperforms the existing methods in terms of higher efficiency. Our approach also provides high scalability, which is suitable for many kinds network address such as IPv4 address, IPv6 address, named data networking address, etc.

Wireless upload services are characterized by rapidly time varying channel conditions and wide size difference in pre-uploaded content generated by the end users. Therefore traditional static cache method used in large content delivery may not apply to wireless upload services. For example, uploading small content to the data center that is closer to the end users especially comparatively well behaved transmission links, caching may impose an extra load on the system that will decrease system efficiency and consume more energy. In this paper, an Adaptive Upload Acceleration Mechanism (AUAM) that can dynamically calculate the Necessity of Caching (NOC) is proposed. NOC is the basic factor to determine whether or not to use caches to accelerate the pre-uploaded content. We tried to find the trade-off between cache access reduction and user experience improvement, which depends on the channel conditions and uploaded data size. In addition, we described the implementation of the AUAM and compared with the traditional static cache method. Through analysis supported by practical upload tests with TD-SCDMA mobile network at China Mobile Labs, it was found that the number of cache access could be reduced to about 53% and the upload speed could be maintained at the same level as traditional static cache method when pre-uploaded content ranges from 10k to 80k bytes.

The fast Internet evolution and rapid development of wireless technologies have made it possible for users to communicate while on the move. Mobile IPv6 (MIPv6) is a candidate solution for next generation mobile Internet. Despite its popularity, MIPv6 still suffers various limitations, e.g., lack of business model and management of enormous and discrete home agents, preventing it from being deployed in large-scale commercial environments. Recently, the ID/Locator split architecture has demonstrated its significant predominance in next generation mobile networks. With the aim of pushing the global deployment of mobility support over IPv6, this study makes an effort to design and evaluate an operational mobility model over IPv6 (OMIPv6) based on ID/Locator split architecture to tackle the problems raised by the current form of MIPv6. In particular, a distributed cloud mobility management system (D-CMMS) is employed to be responsible for maintaining the identification and locations of mobile hosts, as well as providing the name resolution services to the mobile hosts. Furthermore, this paper develops an analytical model considering all possible costs required for the operation of OMIPv6, and adopts it as a costeffective tool to evaluate various costs and operation overheads on the performance of OMIPv6 protocol.

In an effort to solve the problems encountered by the current internet, we have proposed an original hierarchical switched network HSNet architecture with the new address structure and hierarchical switching and controlling protocols including a per-flow load-balancing algorithm for HSNet. To evaluate the effectiveness and accuracy of the proposed architecture and the load-balancing algorithm, this paper conducts extensive NS-2 simulation experiments. Furthermore, we prototype the switching equipment of HSNet based on the Intel IXP2800 Network Processor and test the equipment by virtue of the professional measuring instrument, i.e., Agilent N2X. In addition to the simulation experiments, a test-bed of HSNet is then built based on the developed switching equipments. Both the simulation and real experimental results demonstrate that the per-flow load balancing algorithm can effectively balance the load on network channels for HSNet, and the developed switching equipment can achieve a high throughput.

This paper discusses the optimization and implementation of a new internet resource locating technology-Handle protocol and server. It introduces the basics of Handle protocol and the characteristics of Handle system. Then it focuses on the overall architecture, especially, the design of the Authority Cache and the Dynamic Thread Pool. The innovation of the paper is the proposal of a high performance cache mechanism and the self-adapting thread pool model, with which the latency and the concurrency are balanced. The overall system throughput is improved by all the innovations in this paper, which sets an important foundation for the deployment of Handle system.

This paper proposes an adaptive data aggregation for wireless sensor networks in the TD-SCDMA system. Our design explores a data transmission approach: the atomic MAC layer unit is sent through the TD-SCDMA radio channel by dynamically aggregating one or more sensor packets. A trade-off exists between the desire to reduce header and physical layer overhead by making frames larger, and the need to reduce frame error rates in the noisy channel by using small frame lengths. Clearly the optimum depends on the channel conditions. Our analysis shows that the proposed approach is scalable in maintenance overheads, and performs efficiently in the visible throughput. We implement the proposed protocol in a 3G wireless router and test its feasibility and applicability by a 20-node wireless sensor network. We further conduct extensive simulations to examine the efficiency and scalability of our scheme with varied network settings.

As an important biometric technique, face recognition has received substantial attention from both research communities and the market in the past three decades. However, the poor resolution of captured face images can seriously degrade the performance of conventional FR systems. To this end, many image super-resolution (SR) methods have been proposed recently. In this paper, we regard the face image SR as a kind of image interpolation problem for domain specific image. A missing intensity interpolation method based on Rigid Regression (RR) using a Local Structure Prior (LSP) is presented, which is called RRLSP for short. In order to interpolate missing intensities within a target high-resolution (HR) image, we assume that face image patches at the same position share the similar local structure, which is called LSP, and use the RR to learn the relationship between the low-resolution (LR) pixels and the missing HR pixels of one position patch with the LSP. Performance comparison with existing algorithms shows the effectiveness of the proposed method for face image SR in general.

This paper summarized the development trend of the future Internet testbed, major projects and key technology, and introduced the construction progress of testbed based on research network. This paper gave an overview of the trend of the future Internet testbed and introduced several large projects for the future Internet testbed including PlanetLab, GENI and FIRE. Furthermore, it summarized the virtualization, resource management, instrumentation and measurement, and security which help to construct the future Internet testbed. Also, it discussed network virtualization technology. At the end, it talked about the latest progress of CSTNet Future Internet Testbed.

This paper will focus on the design and implementation of Hierarchically Switched Network switch (based on programmable network hardware) .The HSNET has data plane ,control plane ,and management plane separated which reflects the characteristics of HSNET ,meanwhile ,HSNET is fully compatible with the legacy network and supports the implementation of more innovative architectures which provides the flexibility for the Future Internet and stands for the trend of the development of critial network devices in the Future Internet .

Anycast is a practical solution for service discovery among replication of multiple instances in the Internet. OpenFlow is a good candidate to solve the problems of non-awareness of topology changes, costly address translation and low flexibility resulting from anycast service. This paper presents a new anycast mechanism based on OpenFlow technology using the controller to make anycast routing decisions. The controller possesses the knowledge of network topology and link status to realize load-aware anycasting. The effectiveness and accuracy of the proposed scheme is validated by extensive simulation experiments. The results show that the performance of the presented load-aware anycasting outperforms that of existing anycast schemes.

Multicast transmission can effectively reduce the cost of network resources in delivering packets to a group of receivers. However, it is difficult to adjust the path of the generated multicast spanning tree (MST) if the multicast source and group members are unchanged. OpenFlow has been a key technology to realize the software defined networking and makes the flow control feasible. This paper presents an OpenFlow-based solution to dynamically adjust the path of MST by configuring the paths according to the requirements of the network loads and specific conditions. In addition, we propose a distributed MST strategy to satisfy the requirements of inter-domain multicast transmissions. Experimental results show that the proposed solution can effectively adjust the paths of MST and facilitate the generation of MST by parallel computation in the construction of distributed MST.