Network virtualization is recognized as an enabling technology for the future Internet. It aims to overcome the resistance of the current Internet to architectural change. Application of this technology relies on algorithms that can instantiate virtualized networks on a substrate infrastructure, optimizing the layout for service-relevant metrics. This class of algorithms is commonly known as "Virtual Network Embedding (VNE)" algorithms. This paper presents a survey of current research in the VNE area. Based upon a novel classification scheme for VNE algorithms a taxonomy of current approaches to the VNE problem is provided and opportunities for further research are discussed.

Virtual Network Embedding: A Survey

Due to the default settings of the OSPF, the network takes several tens of seconds to recover from a failure During the convergence, the network service will be disrupted This paper proposed a framework of using OpenFlow to handle the transient link failure OpenFlow separate the data and control plane by shifting the management to a remote centralized controller and the forwarders just forward packets based on flow table We exploit OpenFlow technology into the existing routers The routing information database (RIB) and the flow table construct a two-layer forwarding structure When link failure happens, the controller computed the backup paths for the destinations that affected by the failure link and sent the related flow table to the OpenFlow forwarders All the packets received by the forwarders will be compared against flow table entries first If the packets matched with the flow table, the corresponding actions of the flow entries will be taken Otherwise, the packets will be processed by the RIB After the network recovering from the failure, the controller will remove the temporary flow table used to bypass the failed link in the forwarders and the RIB will take over the packets forwarding Experiments results demonstrate that the proposed framework have a low packet loss to the link failure

A framework of using OpenFlow to handle transient link failure

Increased popularity and wealth-creation potential of conventional mobile phones, smart phones and tablet computers resulted in the unprecedented penetration of wired and wireless communications. Most of such real-time applications require video streaming that needs quality of service (QoS) provisioning in addition to good end-to-end transport performance in network infrastructure. Estimating the reliability of an end-to-end network path is critically important for such applications. Available bandwidth at a node is an important QoS characteristic of the path that is a minimum spare capacity of links constituting a network path. Researchers have proposed various techniques for estimating the bandwidth to increase the network throughput. In the past, bandwidth estimation (BE) techniques developed for wired networks are focused on point-to point dedicated links that are not directly usable in wireless networks. The reason is that these techniques are based on models which are no longer valid in wireless environment. This paper aims to provide a comprehensive survey of the BE techniques proposed till date by researchers in the literature for both wired and wireless networks. We have categorized the BE techniques into four main categories as active probing techniques, passive techniques, techniques only for wireless networks and other BE techniques. A brief outline of each technique is discussed which includes the problem statement, operation methodology, results and applications. Techniques in each category have been compared using various parameters such as accuracy, BE time and overhead.

Survey of Bandwidth Estimation Techniques in Communication Networks

Network virtualization provides more flexibility in network provisioning as it offers physical infrastructure providers (PIP) the possibility of smoothly rolling out many separate networks on top of an existing infrastructure. A major challenge is the embedding problem of mapping virtual networks (VNs) onto PIP infrastructure. In the literature, a good deal of research has focused on providing heuristic approaches to this NP-hard problem, usually with the assumption that the PIP infrastructure is operational at all times. In virtualization environment, a single physical node/link failure can result in one or more logical link failures as it effects all VNs with a mapping that spans over. Setting up a dedicated backup for each VN embedding that is not shared with others is an inefficient use of resources. To address these concerns, this paper proposes two classes of periodic VN protection against link and node failures: (a) in the physical layer, by using a path or segment $p$ -cycle technique and a column generation optimization model, and (b) in the VN layer, by augmenting the topology with redundant resources and subsequently applying a column generation mapping model. Our simulations show a clear advantage of our approaches over benchmarks in terms of PIP profit, backup cost/rate and resource use.

Cost-Efficient Mapping for Fault-Tolerant Virtual Networks

The increase in acceptability and popularity of social media has made extracting information from the data generated on social media an emerging field of research. An important branch of this field is predicting future events using social media data. This paper is focused on predicting box-office revenue of a movie by mining people's intention to purchase a movie ticket, termed purchase intention, from trailer reviews. Movie revenue prediction is important due to risks involved in movie production despite the high cost involved in the production. Previous studies in this domain focus on the use of twitter data and IMDB reviews for the prediction of movies that have already been released. In this paper, we build a model for movie revenue prediction prior to the movie's release using YouTube trailer reviews. Our model consists of novel methods of calculating purchase intention, positive-to-negative sentiment ratio, and like-to-dislike ratio for movie revenue prediction. Our experimental results prove the superiority of our approach compared to three baseline approaches and achieved a relative absolute error of 29.65%.

Movie Revenue Prediction Based on Purchase Intention Mining Using YouTube Trailer Reviews

Network virtualization is viewed as a promising way for overcoming the ossification of the Internet. Virtual network mapping plays an important role in building virtual networks. The substrate network conditions change synchronously when the process of virtual network mapping, when a network failure happens, virtual network mapping without protection may lead to virtual network service interruption. In this paper we proposed an algorithm called RMap which combines virtual network mapping with substrate link backup, this method can guarantee virtual networks resilience and some level of reliability to substrate link failure. Our simulation results show that the proposed algorithm has a good performance.

RMap: An Algorithm of Virtual Network Resilience Mapping

In this article, we discuss various elements contributing to exerting influence on box office in China, which are divided into internal and external factors. Since these factors could merely be quantified by online data sources partially or inaccurately, we propose that relativity analysis is more reasonable than precise revenue prediction. Trailer is selected as the combination of movie content and online behavior prior to releasing. Indexes from seven mainstream video websites are retrieved by the designed big data system which is integrated with the Internet of things technology. Correlation coefficients of different time periods are calculated. We apply multiple linear regression with stepwise method in modeling and prove that watching counts of 1 week before releasing on Youku is the barometer of market performance, especially the first week revenues. We also manifest the power of influential users through constructing Sina Weibo acquisition and analysis system.

https://journals.sagepub.com/doi/pdf/10.1177/1550147716684842

An inside look into the complexity of box-office revenue prediction in China:

Broadcasting has become a basic service that adds to existing DHT systems the ability to search using arbitrary queries as well as disseminate/collection global information. However, the phenomenon of system churn, a continuous process of needs joining and leaving the networks, degrades the broadcast performance of DHT systems and adversely impact the performance of applications built on top of them. In this paper, we present an efficient algorithm for performing a broadcast operation in DHT systems under churn. It is suitable for most DHT systems such as Chord, Tapestry, etc. Through the proposed algorithm, nodes in DHT system could perform broadcast operation rapidly, and will not send any redundant message. Furthermore, it ensures full coverage of the nodes in the system even when the churn rate is high. The algorithm is validated and evaluated through simulations.

An Efficient Broadcast Algorithm in Distributed Hash Table Under Churn

We present a new available bandwidth measurement method based on self-loading periodic streams (SLoPS) in this paper. SLoPS sends probe stream at specific rate and estimates the available bandwidth from the inference of self-induced congestion. Contrary to traditional SLoPS method, the new method infers the congestion from a parameter named interval difference, not end-to-end delays. In probing scheme, the new method exponentially adjust probe rate to demarcate a rough bandwidth range and constringe the rate to available bandwidth value with quick search in the rough range. Since it uses only interval times of packets, our method does not require high time-stamping accuracy nor synchronous at the sender and receiver. We test our method in controlled and reproducible environment using NS2, and the simulation experiments show our method is accurate, quick and non-intrusive.

A New Available Bandwidth Measurement Method Based on Self-Loading Periodic Streams

Reliability evaluation can help to minimize costs and maximize reliability of wireless sensor networks (WSN) To evaluate the reliability of WSN, a mathematical model and an enhanced factoring (EF) algorithm are presented in this paper Being different from traditional factoring algorithm, the node contraction is enhanced by source expansion and the isomorphic sub-networks are identified These operations can help to decrease the number of repeated computations and make the reliability computing fast The traditional factoring algorithm and EF are compared in experiments, and the results show EF is more efficient than traditional factoring for large size network

Xin Li

Papers

RMap: An Algorithm of Virtual Network Resilience Mapping

An inside look into the complexity of box-office revenue prediction in China:

An Efficient Broadcast Algorithm in Distributed Hash Table Under Churn

A New Available Bandwidth Measurement Method Based on Self-Loading Periodic Streams

An Enhanced Factoring Algorithm for Reliability Evaluation of Wireless Sensor Networks