Machine Learning for Networking

Frontmatter

Learning Concave-Convex Profiles of Data Transport over Dedicated Connections

Abstract

Dedicated data transport infrastructures are increasingly being deployed to support distributed big-data and high-performance computing scenarios. These infrastructures employ data transfer nodes that use sophisticated software stacks to support network transport among sites, which often house distributed file and storage systems. Throughput measurements collected over such infrastructures for a range of round trip times (RTTs) reflect the underlying complex end-to-end connections, and have revealed dichotomous throughput profiles as functions of RTT. In particular, concave regions of throughput profiles at lower RTTs indicate near-optimal performance, and convex regions at higher RTTs indicate bottlenecks due to factors such as buffer or credit limits. We present a machine learning method that explicitly infers these concave and convex regions and transitions between them using sigmoid functions. We also provide distribution-free confidence estimates for the generalization error of these concave-convex profile estimates. Throughput profiles for data transfers over 10 Gbps connections with 0–366 ms RTT provide important performance insights, including the near optimality of transfers performed with the XDD tool between XFS filesystems, and the performance limits of wide-area Lustre extensions using LNet routers. A direct application of generic machine learning packages does not adequately highlight these critical performance regions or provide as precise confidence estimates.

Nageswara S. V. Rao, Satyabrata Sen, Zhengchun Liu, Rajkumar Kettimuthu, Ian Foster

Towards Analysing Cooperative Intelligent Transport System Security Data

Abstract

C-ITS (Cooperative Intelligent Transport Systems) provide nowadays a very huge amounts of data from different sources: vehicles, roadside units, operator servers, smartphone applications. These amounts of data can be exploited and analysed in order to extract pertinent information as driver profiles, abnormal driving behaviours, etc. In this paper, we present a methodology for analysis of data provided by a real experimentation of a cooperative intelligent transport system (C-ITS). We have analysed mainly security issues as privacy, authenticity. We have used unsupervised machine learning approaches. The obtained results have shown interesting results in terms of latency, packet delivery ratio.

Brice Leblanc, Emilien Bourdy, Hacène Fouchal, Cyril de Runz, Secil Ercan

Towards a Statistical Approach for User Classification in Twitter

Abstract

In this paper, we propose a novel technique for classifying user accounts on online social networks. The main purpose of our classification is to distinguish the patterns of users from those of organizations and individuals. The ability of distinguishing between the two account types is needed for developing recommendation engines, consumer products opinion mining tools, and information dissemination platforms. However, such a task is non-trivial. Classic and consolidated approaches of text mining use textual features from natural language processing for classification. Nevertheless, such approaches still have some drawbacks like the computational cost and time consumption. In this work, we propose a statistical approach based on post frequency, metadata of user profile, and popularity of posts so as to recognize the type of users without textual content. We performed a set of experiments over a twitter dataset and learn-based algorithms in classification task. Several supervised learning algorithms were tested. We achieved high f-measure results of 96.2% using imbalanced datasets and (GBRT), 1.9% were gains when we used imbalanced datasets with Synthetic Minority Oversampling technique and (RF), this yields 98.1%.

Kheir Eddine Daouadi, Rim Zghal Rebaï, Ikram Amous

RILNET: A Reinforcement Learning Based Load Balancing Approach for Datacenter Networks

Abstract

Modern datacenter networks are facing various challenges, e.g., highly dynamic workloads, congestion, topology asymmetry. ECMP, as a traditional load balancing mechanism which is widely used in today’s datacenters, can balance load poorly and lead to congestion. Variety of load balancing schemes are proposed to address the problems of ECMP. However, these traditional schemes usually make load balancing decision only based on network knowledge for a snapshot or a short time past. In this paper, we propose a Reinforcement Learning (RL) based approach, called RILNET (ReInforcement Learning NETworking), aiming at load balancing for datacenter networks. RILNET employs RL to learn a network and control it based on the learned experience. To achieve a higher granularity of control, RILNET is constructed to route flowlet rather than flows. Moreover, RILNET makes routing decisions for aggregation flows (an aggregation flow is a flow set that includes all flows flowing from the same source edge switch to the same destination edge switch) instead of a single flow. In order to test performance of RILNET, we propose a flow-level simulation and a packet-level simulation, and the both results show that RILNET can balance traffic load much more effectively than ECMP and another load balancing solution, i.e., DRILL. Compared with DRILL, RILNET outperforms DRILL in data loss and maximal link delay. Specifically, the maximal link data loss and the maximal link delay of RILNET are 44.4% and 25.4% smaller than DRILL, respectively.

Qinliang Lin, Zhibo Gong, Qiaoling Wang, Jinlong Li

Building a Wide-Area File Transfer Performance Predictor: An Empirical Study

Abstract

Wide-area data transfer is central to geographically distributed scientific workflows. Faster delivery of data is important for these workflows. Predictability is equally (or even more) important. With the goal of providing a reasonably accurate estimate of data transfer time to improve resource allocation & scheduling for workflows and enable end-to-end data transfer optimization, we apply machine learning methods to develop predictive models for data transfer times over a variety of wide area networks. To build and evaluate these models, we use 201,388 transfers, involving 759 million files totaling 9 PB transferred, over 115 heavily used source-destination pairs (“edges”) between 135 unique endpoints. We evaluate the models for different retraining frequencies and different window size of history data. In the best case, the resulting models have a median prediction error of \(\le \)21% for 50% of the edges, and \(\le \)32% for 75% of the edges. We present a detailed analysis of these results that provides insights into the cause of some of the high errors. We envision that the performance predictor will be informative for scheduling geo-distributed workflows. The insights also suggest obvious directions for both further analysis and transfer service optimization.

Zhengchun Liu, Rajkumar Kettimuthu, Prasanna Balaprakash, Nageswara S. V. Rao, Ian Foster

Advanced Hybrid Technique in Detecting Cloud Web Application’s Attacks

Abstract

Recently cloud computing has emerged the IT world. It eventually promoted the acquisition of resources and services as needed, but it has also instilled fear and user’s renunciations. However, Machine learning processing has proven high robustness in solving security flaws and reducing false alarm rates in detecting attacks. This paper, proposes a hybrid system that does not only labels behaviors based on machine learning algorithms using both misuse and anomaly-detection, but also highlights correlations between network relevant features, speeds up the updating of signatures dictionary and upgrades the analysis of user behavior.

Meryem Amar, Mouad Lemoudden, Bouabid El Ouahidi

Machine-Learned Classifiers for Protocol Selection on a Shared Network

Abstract

Knowledge about the state of a data network can be used to achieve high performance. For example, knowledge about the protocols in use by background traffic might influence which protocol to choose for a new foreground data transfer. Unfortunately, global knowledge can be difficult to obtain in a dynamic distributed system like a wide-area network (WAN).

Therefore, we introduce and evaluate a machine-learning (ML) approach to network performance, called optimization through protocol selection (OPS). Using local round-trip time (RTT) time-series data, a classifier predicts the mix of protocols in current use. Then, a decision process selects the best protocol to use for the new foreground transfer, so as to maximize throughput while maintaining fairness. We show that a protocol oracle would choose TCP-BBR for the new foreground traffic if TCP-BBR is already in use in the background, for proper throughput. Similarly, the protocol oracle would choose TCP-CUBIC for the new foreground traffic if only TCP-CUBIC is in use in the background, for fairness.

Empirically, our k-nearest-neighbour (K-NN) classifier, utilizing dynamic time warping (DTW) measure, results in a protocol decision accuracy of 0.80 for \(k=1\). The OPS approach’s throughput is 4 times higher than that achieved with a suboptimal protocol choice. Furthermore, the OPS approach has a Jain fairness index of 0.96 to 0.99, as compared to a Jain fairness of 0.60 to 0.62, if a suboptimal protocol is selected.

Hamidreza Anvari, Jesse Huard, Paul Lu

Common Structures in Resource Management as Driver for Reinforcement Learning: A Survey and Research Tracks

Abstract

In the era of growing digitalization, dynamic resource management becomes one of the critical problems in many application fields where, due to the permanently evolving environment, the trade-off between cost and system performance needs to be continuously adapted. While traditional approaches based on prior system specification or model learning are challenged by the complexity and the dynamicity of these systems, a new paradigm of learning in interaction brings a strong promise - based on the toolset of model-free Reinforcement Learning (RL) and its great success stories in various domains. However, current RL methods still struggle to learn rapidly in incremental, online settings, which is a barrier to deal with many practical problems. To address the slow convergence issue, one approach consists in exploiting the system’s structural properties, instead of acting in full model-free mode. In this paper, we review the existing resource management systems and unveil their common structural properties. We propose a meta-model and discuss the tracks on how these properties can enhance general purpose RL algorithms.

Yue Jin, Dimitre Kostadinov, Makram Bouzid, Armen Aghasaryan

Inverse Kinematics Using Arduino and Unity for People with Motor Skill Limitations

Abstract

In this document, the creation process of an application is detailed that can use various sensors connected and managed by an Arduino UNO board to capture the movement from the extremities in people with limited movement. It is noteworthy to mention that the results of said application are not discussed here but only the creation process is described. The first part is dedicated to describing the hardware required to use the app’s programmed technology. A brief overview of the Arduino platform is also given followed by a description of the sensors used for calculation and capture of people’s movement. The selection of the sensors is justified, and their operation is presented. The second part focuses on the construction of the application starting by offering a synthesized view of the Unity platform up to the development process. Additionally, the basic concepts to generate the 3D models are explained with the purpose of allowing anyone that reads this document to replicate the project in a simple manner.

Nicolás Viveros Barrera, Octavio José Salcedo Parra, Lewys Correa Sánchez

Delmu: A Deep Learning Approach to Maximising the Utility of Virtualised Millimetre-Wave Backhauls

Abstract

Advances in network programmability enable operators to ‘slice’ the physical infrastructure into independent logical networks. By this approach, each network slice aims to accommodate the demands of increasingly diverse services. However, precise allocation of resources to slices across future 5G millimetre-wave backhaul networks, to optimise the total network utility, is challenging. This is because the performance of different services often depends on conflicting requirements, including bandwidth, sensitivity to delay, or the monetary value of the traffic incurred. In this paper, we put forward a general rate utility framework for slicing mm-wave backhaul links, encompassing all known types of service utilities, i.e. logarithmic, sigmoid, polynomial, and linear. We then introduce Delmu, a deep learning solution that tackles the complexity of optimising non-convex objective functions built upon arbitrary combinations of such utilities. Specifically, by employing a stack of convolutional blocks, Delmu can learn correlations between traffic demands and achievable optimal rate assignments. We further regulate the inferences made by the neural network through a simple ‘sanity check’ routine, which guarantees both flow rate admissibility within the network’s capacity region and minimum service levels. The proposed method can be trained within minutes, following which it computes rate allocations that match those obtained with state-of-the-art global optimisation algorithms, yet orders of magnitude faster. This confirms the applicability of Delmu to highly dynamic traffic regimes and we demonstrate up to 62% network utility gains over a baseline greedy approach.

Rui Li, Chaoyun Zhang, Pan Cao, Paul Patras, John S. Thompson

Malware Detection System Based on an In-Depth Analysis of the Portable Executable Headers

Abstract

Malware still pose a major threat for cyberspace security. Therefore, effective and fast detection of this threat has become an important issue in the security field. In this paper, we propose a fast and highly accurate detection system of Portable Executable (PE) malware. The proposed system relies on analyzing the fields of the PE-headers using a basic way and a more in-depth way in order to generate a set of standard attributes (SAT), and meaningful attributes (MAT) respectively. The decision phase is conducted by leveraging several machine learning classifiers, which are trained using the best K attributes according to two different feature selection methods. The experimental results are very promising, as our system outperforms two state-of-the-art solutions with respect to detection accuracy. It achieves an accuracy of 99.1% and 100% using 10-folds cross validation and train-test split validation, respectively. In both validation approaches, we only use less than 1% out of the initial set of 1329 extracted attributes. Also, our system is able to analyze a file in 0.257 s.

Mohamed Belaoued, Bouchra Guelib, Yasmine Bounaas, Abdelouahid Derhab, Mahmoud Boufaida

DNS Traffic Forecasting Using Deep Neural Networks

Abstract

With the continuous growth of Internet usage, the importance of DNS has also increased, and the large amount of data collected by DNS servers from users’ queries becomes a very valuable data source, since it reveals user patterns and how their Internet usage changes through time. The periodicity in human behavior is also reflected in how users use the Internet and therefore in the DNS queries they generate. Thus, in this paper we propose the use of Machine Learning models in order to capture these Internet usage patterns for predicting DNS traffic, which has a huge relevance since a big difference between the expected DNS traffic and the real one, could be a sign of an anomaly in the data stream caused by an attack or a failure. To the best of the authors’ knowledge this is the first attempt of forecasting DNS traffic using Neural Networks models, in order to propose an unsupervised and lightweight method to perform fast detection of anomalies in DNS data streams observed in DNS servers.

Diego Madariaga, Martín Panza, Javier Bustos-Jiménez

Energy-Based Connected Dominating Set for Data Aggregation for Intelligent Wireless Sensor Networks

Abstract

The main mission of deploying sensors is data collection and the main sensor resource to save is energy. For this reason, data aggregation is an important method to maximize sensors’ lifetime. Aggregating sensed data from multiple sensors eliminates the redundant transmissions and provides fused information to the sink. It has been proved in the literature that a structure based data aggregation gives better results in terms of packet delivery and energy saving which prolong the network lifetime. In this paper, we propose a novel approach called Distributed Connected Dominating Set for Data Aggregation (DCDSDA) to construct our network topology. The sensors of the network compute in a distributed way and based on the residual energy of each sensor, a connected dominating set to form a virtual backbone. This backbone forms a tree topology and as it is computed and maintained in a distributed way based on predefined energy constraints, it represents an intelligent fault tolerance mechanism to maintain our network and to deal with packet loss. The simulation results show that our proposed method outperforms existing methods.

Basem Abid, Sarra Messai, Hamida Seba

Touchless Recognition of Hand Gesture Digits and English Characters Using Convolutional Neural Networks

Abstract

Computer technology has changed the way humans define success. Gesture recognition is a new and intuitive interaction method, as it does not require any mechanical interface for interaction. The model proposed takes input from the user using an infrared emitter and a web camera and this is used to generate a 28\(\,\times \,\)28 image contains the drawn character. This is passed through a convolutional neural network which detects the character drawn. The benefits include reliability and cost-effective nature of the proposed system. Such an approach can help people with disabilities.

Ujjwal Peshin, Tanay Jha, Shivam Kantival

LSTM Recurrent Neural Network (RNN) for Anomaly Detection in Cellular Mobile Networks

Abstract

Anomaly detection can show significant behavior changes in the cellular mobile network. It can explain much important missing information and which can be monitored using advanced AI (Artificial Intelligent) applications/tools. In this paper, we have proposed LSTM (Long Short-Term Memory) based RNN (Recurrent Neural Network) which can model a time series profile for LTE network based on cell KPI values. We have shown in this paper that the dynamic behavior of a single cell can be simplified using a combination of a set for neighbor cells. We can predict the profile and anomalous behavior using this method. According to the best of our knowledge this approach is applied here for the first time for cell level performance profile generation and anomaly detection. In a related work, they have proposed ensemble method to compare different KPIs and cell performance using machine learning algorithm. We have applied DNN (Deep Neural Network) to generate a profile on KPI features from historical data. It gave us deeper insight into how the cell is performing over time and can connect with the root causes or hidden fault of a major failure in the cellular network.

S. M. Abdullah Al Mamun, Mehmet Beyaz

Towards a Better Compromise Between Shallow and Deep CNN for Binary Classification Problems of Unstructured Data

Abstract

Deep Neural Network is a large scale neural network. Deep Learning, refers to training very large Neural Networks in order to discover good representations, at multiple levels, with higher-level learned features. The rise of deep learning is especially due to the technological evolution and huge amounts of data. Since that, it becomes a powerful tool that everyone can use specifically on supervised learning, because it’s by far the dominant form of deep learning today. Many works based on Deep learning have already been proposed. However, these works have not given any explanation on the choice of the number of the network layers. This makes it difficult to decide on the appropriate deep of the network and its performances for a specific classification problem. In this paper the objective is threefold. The first objective was to study the effect of facial expressions on facial features deformations and its consequences on gender recognition. The second objective is to evaluate the use of Deep learning in the form of transfer learning for binary classification on small datasets (containing images with different Facial expressions). Our third goal is then to find a compromise between too much capacity and not enough capacity of the used deep Neural Network in order to don’t over fit nor under fit. Three different architectures were tested: a shallow convolutional neural network (CNN) with 6 layers, a deep CNN VGG16 (16 layers) and very deep CNN RESNET50 (50 Layers). Many conclusions have been drawn.

Khadoudja Ghanem

Reinforcement Learning Based Routing Protocols Analysis for Mobile Ad-Hoc Networks

Global Routing Versus Local Routing

Abstract

Energy consumption and maximize lifetime routing in Mobile Ad hoc Network (MANETs) is one of the most important issues.

In our paper, we compare a global routing approach with a local routing approach both using reinforcement learning to maximize lifetime routing.

We first propose a global routing algorithm based on reinforcement learning algorithm called Q-learning then we compare his results with a local routing algorithm called AODV-SARSA.

Average delivery ratio, End to end delay and Time to Half Energy Depletion are used like metrics to compare both approach.

Redha Mili, Salim Chikhi

Deep Neural Ranking for Crowdsourced Geopolitical Event Forecasting

Abstract

There are many examples of “wisdom of the crowd” effects in which the large number of participants imparts confidence in the collective judgment of the crowd. But how do we form an aggregated judgment when the size of the crowd is limited? Whose judgments do we include, and whose do we accord the most weight? This paper considers this problem in the context of geopolitical event forecasting, where volunteer analysts are queried to give their expertise, confidence, and predictions about the outcome of an event. We develop a forecast aggregation model that integrates topical information about a question, meta-data about a pair of forecasters, and their predictions in a deep siamese neural network that decides which forecasters’ predictions are more likely to be close to the correct response. A ranking of the forecasters is induced from a tournament of pair-wise forecaster comparisons, with the ranking used to create an aggregate forecast. Preliminary results find the aggregate prediction of the best forecasters ranked by our deep siamese network model consistently beats typical aggregation techniques by Brier score.

Giuseppe Nebbione, Derek Doran, Srikanth Nadella, Brandon Minnery

The Comment of BBS: How Investor Sentiment Affects a Share Market of China

Abstract

This paper studies the influence of investor sentiment on the A share market of China. According to the behavioral financial theory, transformation of investor sentiment will trigger irrational transaction behavior and have an influence on the Chinese stock market. We study the effect of more than 23 million investor’s comments posted on EastMoney.​com, which is the biggest stock BBS in China. We utilize TextCNN to mine emotional tendency of investor comment stock comment, classify comments into positive, negative and neutral. The classified accuracy of validate set can reach 90%. And utilize such emotional tendency to define investor sentiment index. Based on our research, we find that the correlation between sentiment index and Shanghai Composite Index (SHCI) is positive, statistically significant and exponentially decays in period of time. Besides, with Hurst parameter H, it indicates that investor sentiment have long-range correlations, investor sentiment will develop as the current trend.

Xuanlong Weng, Yin Luo, Jianbo Gao, Haishan Feng, Ke Huang

A Hybrid Neural Network Approach for Lung Cancer Classification with Gene Expression Dataset and Prior Biological Knowledge

Abstract

Lung cancer has continued to be the leading cause of related mortality and its frequency is rising daily worldwide. A reliable and accurate classification is essential for successful lung cancer diagnosis and treatment. Gene expression microarray, which is a high-throughput platform, makes it possible to discover genomic biomarkers for cancer diagnosis and prognosis. This study proposes a new approach of using improved Particle Swarm Optimization (IMPSO) technique to improve the Multi-Layer Perceptrons (MLP) neural network prediction accuracy. The MLP weights and biases are computed by the IMPSO for more accurate lung cancer prediction. The proposed discriminant method (MLP-IMPSO) integrates the prior knowledge of lung cancer classification on the basis of gene expression data to enhance the classification accuracy. Evaluations and comparisons of prediction performance were thoroughly carried out between the proposed model and the representative machine learning methods (support vector machine, MLP, radial basis function neural network, C4.5, and Naive Bayes) on real microarray lung cancer datasets. The cross-data set validations made the assessment reliable. The performance of the proposed approach was better upon the incorporation of prior knowledge. We succeeded in demonstrating that our method improves lung cancer diagnosis accuracy with prior biological knowledge. The evaluation results also showed the effectiveness the proposed approach for lung cancer diagnosis.

Hasseeb Azzawi, Jingyu Hou, Russul Alanni, Yong Xiang

Plant Leaf Disease Detection and Classification Using Particle Swarm Optimization

Abstract

The loss of crops due to diseases is a major danger to food security. It is important to develop the requisite infrastructure and tools for the detection of diseases in crops. The opportunity to detect diseases in crops has increased manifolds with the rise in the number of smartphone users and improved network connectivity. In this paper, we provide an approach to detect and classify plant leaf diseases. The methodology involves image acquisition, pre-processing of the images, feature extraction followed by feature selection and finally the classification of plant diseases. A deep convolutional neural network was trained to extract features from the input image. An optimal set of features is selected using Particle Swarm Optimization (PSO) and are classified into 23 different classes, including both healthy and diseased categories. Apropos, by employing this technique, the plant leaf images are classified with an accuracy of 97.39%.

Rishabh Yadav, Yogesh Kumar Rana, Sushama Nagpal

Game Theory Model for Intrusion Prevention in Computer Networks

Abstract

No system is completely secure, but it’s possible to analyze system vulnerabilities, take security measures and thus reduce the likelihood of system intrusions. One tool to reduce this probability of intrusion is to implement intrusion prevention systems, which analyze network traffic to determine anomalous behavior and take security measures with the agents that exhibit these behaviors. This document proposes a game theory model based in an evaluation of works related in the area that could be the basis of an algorithm for the prevention of intrusions in networks.

Julián Francisco Mojica Sánchez, Octavio José Salcedo Parra, Miguel J. Espitia R.

A Survey: WSN Heterogeneous Architecture Platform for IoT

Abstract

Internet of Things (IoT) is a novel paradigm that allows millions of smart devices to be connected to the Internet. Such devices can be sensors/actuators, which are able to operate and transmit data to other systems in an autonomous way. In fact, IoT is about autonomous and heterogeneous devices, data and connectivity. As millions of devices are connected, internet of things will require more improvement in terms of platform deployment. In this context, IoT must coexists with several other paradigms (Cloud-Computing, Big-data, SDN…) in other to satisfy its new features. Next to that, Wireless Sensor Networks (WSNs) are one of the most important components in IoT.

This paper is the result of a state of the art analysis and performance evaluation of several SDN architectures designed for the Internet of Things platforms. At the end of the paper, we present an overview of our architecture based on SDN and SDR methods for WSN as a part of the Internet of Things.

Naila Bouchemal, Sondes Kallel, Nardjes Bouchemal

An IoT Framework for Detecting Movement Within Indoor Environments

Abstract

Tracking people indoors can be valuable in smart living scenarios such as tracking shoppers in a mall or in healthcare situations when tracking the movement of elderly patients can allow them to remain more independent. Determining accurate movement of people indoors is problematic however as there is no universal tracking system such as GPS which works indoors. Instead, a range of techniques are used based on technologies such as cameras, radio frequency identification, WiFi, Bluetooth, pressure pads and radar are used to track people and objects within indoor environments. The most common technologies for tracking are Bluetooth and WiFi. Many Internet of Things (IoT) devices support these protocols and can therefore act as beacons and hubs for movement detection indoor. We provide here an overview of an IoT focused framework which allows the plug and play of Bluetooth and WiFi devices in addition to integrating passive and active approaches to determining the movement of people indoors.

Kevin Curran, Gary Mansell, Jack Curran

A Hybrid Architecture for Cooperative UAV and USV Swarm Vehicles

Abstract

This paper is interested in the problem of monitoring and cleaning dirty zones of oceans, dealing with the notion of path planning for semi-autonomous unmanned vehicles. We present a hybrid cooperative architecture for unmanned aerial vehicle (UAV) to monitor ocean region and clean dirty zones with the help of swarm unmanned surface vehicles (USVs). In the path planning problem, unmanned vehicles must plan their path from the starting to the goal position. In this article, we propose a solution to handle the problem of trajectory planning for semi-autonomous cleaning vehicles. This solution is based on the proposed Genetic Algorithm (GA). In order to optimize this process, our proposed solution detects and reduces the pollution level of the ocean zones while taking into account the problem of fault tolerance related to these vehicles.

Salima Bella, Assia Belbachir, Ghalem Belalem

Detecting Suspicious Transactions in IoT Blockchains for Smart Living Spaces

Abstract

The idea of connecting physical things and cyber components to enable new and richer interactions is a key component in any smart space concept. One of the central challenges in these new smart spaces is the access control of data, services and things. In recent years, Distributed Ledger technology (DLT) like Blockchain Technology (BCT), emerged as the most promising solution for decentralized access management. Using capability-based access control, access to data/services/things is achieved by transferring tokens between the accounts of a distributed ledger. Managing how the access tokens are transferred is, of course, a major challenge. Within the IoT space, smart contracts are at the center of most of the proposals for DLT/BCT networks targeting access control. The main problem in using smart contracts as a means for checking if and what access token can be transferred from one account to another is their immutability and accessibility. Smart contracts and chain code are by design meant to be immutable since they represent a binding contract between parties. In addition, they need to be accessible since they are to be executed on many nodes. This allows an attacker to study them and design the attack in a manner that passes the rules of the smart contract/chain code. This paper focuses on the use of metadata as a more effective means to prevent attackers from gaining access to data/services/things in a smart living space.

Mayra Samaniego, Ralph Deters

Intelligent ERP Based Multi Agent Systems and Cloud Computing

Abstract

The use of agents makes ERP systems more intelligent. They provide capabilities to independently interact with their surrounding environment and performing autonomous actions while cooperating with other systems Regarding the cloud computing, its use is almost obvious. It provides access to full-function applications at a rational cost without a substantial upfront for hardware and software spending. In this paper, we propose an Intelligent ERP system based on new technologies: agents and cloud computing.

This combination makes ERP users (managers, customers, employees, etc.) expect data to be made available to them via the most widely used communication mobile device.

Nardjes Bouchemal, Naila Bouchemal

Backmatter