Download e-book Gateway Portals to the City: Infrastructure for Sustainable Urbanization

Free download. Book file PDF easily for everyone and every device. You can download and read online Gateway Portals to the City: Infrastructure for Sustainable Urbanization file PDF Book only if you are registered here. And also you can download or read online all Book PDF file that related with Gateway Portals to the City: Infrastructure for Sustainable Urbanization book. Happy reading Gateway Portals to the City: Infrastructure for Sustainable Urbanization Bookeveryone. Download file Free Book PDF Gateway Portals to the City: Infrastructure for Sustainable Urbanization at Complete PDF Library. This Book have some digital formats such us :paperbook, ebook, kindle, epub, fb2 and another formats. Here is The CompletePDF Book Library. It's free to register here to get Book file PDF Gateway Portals to the City: Infrastructure for Sustainable Urbanization Pocket Guide.
This book is a record of the proceedings of a conference on gateway portals to cities from the perspective of sustainable urbanization. It was held on 19 May.
Table of contents

Grid Portals Documents. PwC: Insurance Portals - gateways to growth Technology. Introduction to broker portals Technology. Portals Numerology Documents. Introduction to Competitive Intelligence Portals Business. Enterprise Portals - Gateway to the Gold Technology. Library Portals Documents.

Urbanization and migration

From Portals to platforms Education. Portals feina Documents. Wanadoo Portals Documents. Portals to the Past Documents. Strategies to Designing Beautiful Portals Technology.

News and Views from the Global South

As the main goal is to keep the gateway as simple as possible, thus our gateway is equipped with an IEEE The management of the virtualization layer is performed by OpenStack Footnote 5. We use the Open vSwitch to classify, to tag the ingress flows and to forward them to the correct function chain in the NFVI.

However, the implementations available for the NSH protocol are still initial and the performance achieved is not satisfactory [ 16 ]. Therefore, a feasible alternative to establishing a coherent and performing service path is the usage of well-established tunneling protocols, such as GRE and VXLAN, to create the service function chains. In this paper, we use the GRE protocol to implement the chaining of service functions. The environment configuration is organized with a node acting as an OpenStack cloud management controller and an OpenDaylight software-defined network controller Footnote 6 ; the three other nodes are dedicated to processing and memory virtualization, through Linux KVM Footnote 7 , and to distributed disk storage through Ceph Footnote 8.

In this work, we implement the virtual network functions as virtual machines running Linux Ubuntu In order to evaluate the effectiveness of our infrastructure to offer real network services, we have defined and developed two virtual network functions to provide security and QoS services to IoT applications. The first VNF classifies the traffic of each IoT domain between legitimate or malicious, using machine learning algorithms. The second enforces quality of service policies on traffic by forwarding flows by queues with previously allocated resources. Flow forwarding through correct queues is performed by rules on a software switch instantiated as a virtual network function.

We evaluate the performance of our IoT infrastructure in regard of the overhead of the access gateway and the effectiveness of the proposed virtual network function. We have defined two main scenarios to evaluate our infrastructure. In the first one, we aim at assessing the performance and the overhead introduced by virtualizing multiple wireless interfaces at the same access point, which provide domain isolation to different IoT applications. In this scenario, to emulate the traffic of several IoT devices, we use two portable computers equipped with Intel Core iM processor, 6 GB of RAM and built-in wireless network interface cards Footnote In the second scenario, we analyze the two proposed virtual network functions, previously described.

Thus, first we compare the performance of three classification algorithms based on machine learning to identify malicious traffic and legitimate traffic. Next, we evaluate the effectiveness of the proposed network virtualization infrastructure to protect the access network of IoT devices from attacks; and to provide quality of service to priority traffic. In the first experiment, we verify the performance of the connection between the network of IoT devices and the network virtualization infrastructure in emulating different delay values for the connection between the gateway and the NFVI.

Therefore, this experiment measures the maximum rate of packets sent by IoT devices and received by the data consumers. It is worth mentioning that in our prototype the maximum transfer unit MTU was set to B due to overloads with encapsulations. Traffic generation was accomplished by creating UDP packet flows at constant rates of packets per second.

In Fig. Another interesting result is that the delay between the gateway and the NFVI has little influence on the packet arrival rate at the NFVI, as the rate of received packets remains almost the same while the delay varies from 0 to ms. The results reinforce the idea of placing a gateway away from the NFVI does not affect the capability of the NFVI to process and to forward almost the same packet rate.

When using a larger packet size, B, the limiting factor is the transmission rate achieved by the gateway wireless network card. Checking the actual baud rate achieved by the network card, set to operate in the IEEE The delay between the gateway and the network virtualization infrastructure does not impact the achieved rates.

The experiments were performed with UDP flows with a small packets of 64 bytes and b large packets of bytes. Next we evaluate the virtualization overhead of the wireless network interface.

Gateway Portals and the Quest for Sustainable Urbanization | Inter Press Service

The evaluation considers four simple scenarios for applying the virtualization infrastructure to IoT. It is important to highlight that we use laptops to emulate the traffic of several IoT devices. In the first scenario, a computer is connected to a virtual access point c 1 ap 1 ; in the second, two computers connect to a single virtual access point c 2 ap 1 ; the third, a computer joins one of two virtual access points c 1 ap 2 ; and finally, in the last scenario, two computers connect to two distinct virtual access points c 2 ap 2.

Clearly, when no delay is added 0 ms , the impact of the virtualization delay is more important but is limited to 10 ms. As the connection delay between the gateway and the NFVI increases, the impact of wireless access virtualization is smoothed. Virtualization of the wireless link. Evaluation of scenarios with one computer and a one single virtual access point 1c1ap ; two computers and one access point 2c1ap ; two computers and one access point 2c1ap ; and two computers and two access points 2c2ap.

In ‘The Future of Cities,’ innovative responses to urban issues

We also measured the aggregate bandwidth reached by IoT devices in each scenario, Fig. The aggregated bandwidth is not affected by the wireless network virtualization for B packets. First, we observe that with just one computer, adding a virtual AP does not impact the aggregate bandwidth. But most interesting, when we compare the scenarios with two computers, each one connecting to an isolated virtual AP, the aggregate bandwidth increases, instead of two computers sharing a single AP.

The aggregate bandwidth increases in the scenario with two access points and two computers. Such behavior is due to the fact that when performing the virtualization, the scheduling of wireless nodes is performed by the kernel of the gateway operating system instead of performing the scheduling by the controller of the wireless network card.

The wireless network card features a controller with low processing power since it is a low-cost wireless card.

Can a city’s downtown be its portal to prosperity?

As mentioned above, we have deployed two virtual network functions as use cases to evaluate the effectiveness of our infrastructure in providing complex network services for IoT applications. In this scenario, we consider that IoT devices shares their access network with other wireless users, smartphones, and devices with Internet access, such as an IP camera. We also assume that all IoT devices are susceptible to attacks and malware infection.


  • Gateway Portals to the City: Infrastructure for Sustainable Urbanization - World Peace Ever TV.
  • The Devils cycle;
  • Search form!
  • Data on urbanization and migration.
  • Search stock photos by tags.

Thus, a possible network protection is the instantiation of a virtual network function able to classify legitimate traffic and malicious traffic, and subsequently to enforce conformance policies to the traffic. In the first experiment, we compare the performance of three classification algorithms based on machine learning to identify malicious traffic and legitimate traffic from devices connected to the IoT network.

The traffic classification between legitimate and malicious flows depends on the training and evaluation of classification algorithms. For this purpose, a training and testing dataset, composed of legitimate data and labeled attack data, was created Footnote Particularly, the cameras were accessed to generate video streaming data, accesses to FTP servers for video and photo transfer, and date synchronization via NTP. The attack data were obtained from the data collected by Garcia et al. The dataset consists of flows identified by a tuple composed of the source and destination IP address, source and destination transport ports, and transport protocol [ 25 ].

Our compiled dataset counts with traffic from 15 devices both from captured laboratory usage traffic and the botnet dataset. The features that were used for generating the dataset are a subset of the numerical features provided by the flowtbag Footnote 14 network characterization application. We also added some tagging features which indicates whether the flow comes from one of the top 10 most accessed services, in terms of the number of flows. We added 10 new features into the dataset, that marks if the flow belongs to the specific service. It is important to use tagging features to keep some information about the service whereas we still are able to calculate correlation and to apply the principal components analysis over the dataset [ 25 ].

We applied the algorithms for all features and for the scenarios in which the dimensionality of the problem was reduced using Principal Component Analysis PCA. Note that classification using boosted decision trees with the gradient boosting algorithm presents the best accuracy when compared to neural networks. The relation between true positive and false positive rates of the classification algorithms, the ROC curve Footnote 15 shown in Fig.

The nominal accuracy was 0. Evaluation of machine learning algorithms for the classification of IoT traffic when using PCA for dimensionality reduction and when using all the features.

The Decision Tree algorithm is the one that presents the best trade-off between true positives and false positives. In the next experiment, we evaluate the ability of our virtualization infrastructure at providing quality of service QoS , taking countermeasures and applying policies to packet flows. For this purpose, we use two notebooks connected to the wireless network to emulate the traffic of two sets of IoT devices with different QoS requirements when accessing the network. Service differentiation is provided by a VNF that re-directs the flows from IoT devices to distinct queues, which are implemented in a software switch Open vSwitch.

During the experiment, both nodes execute a UDP flow at their maximum rate. Flow differentiation in a IoT scenario. In 30 s, the bandwidth controller is activated. The last experiment aims at evaluating the performance of the proposed infrastructure under a Denial of Service DoS attack. Hence, we combine the two virtual network functions to protect IoT applications from DoS attacks. It is worthy to note that our queue-based approach differs from classical approach since our proposal performs traffic policy enforcement in a virtual network function running on a centralized NFVI, unlike other proposals that enforce traffic policies on the access gateways.

This difference where applying the queue limits enables the outsourcing of gateway functions to the NFVI but implies an indirect bandwidth control on the access gateway. Therefore, we avoid the degradation of the service provided by the camera. The protection provided by our infrastructure is accomplished due to the existence of a special gateway that virtualizes the wireless network, and the availability of processing and memory isolated resources in the infrastructure to handle the amount of traffic of the DoS attacker.

It is worth mentioning that our proposal is able to correctly classify the malicious traffic in real-time, which allows redirecting it to the service function chain where it can be shaped with a strict queue policy. Petrolo et al. The authors define the concept of Cloud of Things CoT that consists of using cloud computing environments to provide a platform for integrating data silos of IoT. Santana et al. Thus, Santana et al. Zhang et al.

Colleges, School and Institutes

The authors argue that it is necessary to provide a management and data processing architecture aware of data security and privacy requirements. Atzori et al. The authors point out that RFID systems are small, low cost and energy is not a limiting factor. Wireless sensor networks have high radio coverage and communication does not require the presence of a reader, while RFID, reader and sensor systems are asymmetric.

RFID sensor networks enable detection, computation, and communication in a passive system. In turn, Adelantado et al. Thus, each different access network for the IoT devices presents distinct characteristics and network requirements. Quin et al.