Abdolreza Abhari

Workload generation for YouTube

This paper introduces a workload characterization study of the most popular short video sharing service of Web 2.0, YouTube. Based on a vast amount of data gathered in a five-month period, we analyzed characteristics of around 250,000 YouTube popular and regular videos. In particular, we collected lists of related videos for each video clip recursively and analyzed their statistical behavior. Understanding YouTube traffic and similar Web 2.0 video sharing sites is crucial to develop synthetic workload generators. Workload simulators are required for evaluating the methods addressing the problems of high bandwidth usage and scalability of Web 2.0 sites such as YouTube. The distribution models, in particular Zipf-like behavior of YouTube popular video files suggests proxy caching of YouTube popular videos can reduce network traffic and increase scalability of YouTube Web site. YouTube workload characteristics provided in this work enabled us to develop a workload generator to evaluate the effectiveness of this approach.

Cluster-discovery of Twitter messages for event detection and trending

Abstract Social media data carries abundant hidden occurrences of real-time events. In this paper, a novel methodology is proposed for detecting and trending events from tweet clusters that are discovered by using locality sensitive hashing (LSH) technique. Key challenges include: (1) construction of dictionary using incremental term frequency–inverse document frequency (TF–IDF) in high-dimensional data to create tweet feature vector, (2) leveraging LSH to find truly interesting events, (3) trending the behavior of event based on time, geo-locations and cluster size, and (4) speed-up the cluster-discovery process while retaining the cluster quality. Experiments are conducted for a specific event and the clusters discovered using LSH and K-means are compared with group average agglomerative clustering technique.

Web object-based storage management in proxy caches

Proxy caches are essential to improve the performance of the World Wide Web and to enhance user perceived latency. Appropriate cache management strategies are crucial to achieve these goals. In our previous work, we have introduced Web object-based caching policies. A Web object consists of the main HTML page and all of its constituent embedded files. Our studies have shown that these policies improve proxy cache performance substantially. In this paper, we propose a new Web object-based policy to manage the storage system of a proxy cache. We propose two techniques to improve the storage system performance. The first technique is concerned with prefetching the related files belonging to a Web object, from the disk to main memory. This prefetching improves performance as most of the files can be provided from the main memory rather than from the proxy disk. The second technique stores the Web object members in contiguous disk blocks in order to reduce the disk access time. We used trace-driven simulations to study the performance improvements one can obtain with these two techniques. Our results show that the first technique by itself provides up to 50% reduction in hit latency, which is the delay involved in providing a hit document by the proxy. An additional 5% improvement can be obtained by incorporating the second technique.

A Weighted Energy Efficient Clustering (WEEC) for Wireless Sensor Networks

Wireless sensor networks consist of hundreds or thousands of nodes and are broadly used for collecting data from the environment. In this type of network nodes communicate with each other and send their data to the base station. Since sensor nodes are battery limited, it is crucial to minimize the amount of energy they dissipate for communication. The LEACH protocol is an elegant clustering algorithm for sending information to base station. It uses clustering, therefore each node only communicates with their cluster head nodes and then those cluster heads communicate with base station. In this paper we improved the LEACH algorithm and proposed a weighted energy efficient clustering algorithm (WEEC) algorithm for wireless sensor networks. We take into consideration the location of each node while clusters are forming. The simulation result proves that our proposed scheme noticeably increases the life time of the network.

HTTP modification to reduce client latency

In this paper, we have developed an additional enhancement to the HTTP protocol. This modification eliminates the need for the client to wait for the delivery of the HTML file before being able to request the embedded objects. In our solution, the server delivers the embedded objects of a Web page to the client without being explicitly requested. In essence, the requests for the embedded objects are generated by the server instead of by the client. We introduced, in the form of a module for Apache HTTP server, a mechanism for preloading Web page from server to client. The client Web browser is able to receive this modified response through the use of a custom built proxy.

Modeling of multimedia files on the Web 2.0

In this paper we introduced a workload characterization study of the most popular video sharing service, YouTube, on the Web 2.0. For approximately a two-month period we collected the information of more than 17,000 video files and investigated the file attributes and popularity characteristics of YouTube. Since YouTube is considered as a huge video library, its access pattern has an important impact on the Internet traffic distribution. Comprehension of YouTube features and similar video sharing sites is critical to analyze network traffic and to develop novel user generated contents (UGC) services. This distribution models especially Zipf-like behavior of popular video files suggests caching may reduce network traffic and increase scalability of YouTube Web site.

Improving the delivery of multimedia embedded in web pages

The purpose of this work is to reduce the delivery time for the initial portion of multimedia objects over HTTP protocol. The multimedia files we consider are those embedded in web pages. By prefetching embedded media from server to client, we are able to overcome the effects of network latency. We implement prefetching in a custom built HTTP server capable of anticipating future requests from the client, and delivering data without it explicitly being requested. To allow the client to receive files not requested, we use a custom built proxy, to run on the client system. Our custom server and proxy implement modifications to the HTTP protocol to allow multiple files to be delivered in a single transmission.

Characterization of user networks in Facebook

Network service providers have been faced with some technical challenges that are presented by Facebook and other online social networking technologies. Therefore, it is better to know more about characteristics of these networks. There are several interesting topics worthy of conducting research in Facebook. One of the research directions is analyzing user networks in Facebook. The other research direction is analyzing fan networks of Facebook web pages. We have pursued both research directions in this paper. In this paper, we observed that the user network degree distribution in Facebook follows a lognormal distribution. This paper also presents that the distribution of increase in the number of fans of the web pages follows the Weilbull distribution.

Enhancing broadcast authentication in sensor networks

Due to the nature of wireless sensor networks, security is a critical problem that needs to be further researched and developed. Resource constrained and usually unattended sensors are much vulnerable to malicious attackers that may impersonate the senders by altering broadcast messages. Authenticating received messages is a crucial matter that needs to be investigated closely in this regard. Recently proposed TESLA based techniques have embarked on resolving the authentication problem by employing symmetric encryption and achieving the desired security level by mimicking asymmetric encryption through delayed key disclosure. The suggested delay renders the network vulnerable to Denial of Service attack since an adversary can flood the nodes by sending bogus messages and forcing the sensors to buffer the messages until they receive the corresponding delayed keys. Several novel techniques have been proposed to achieve immediate authentication in TESLA methods to alleviate this threat. In the process, other factors such as reliability, security and buffer requirements may have been compromised which need careful consideration. In this paper a Low Buffer μ Tesla protocol which has been presented in [1] is adapted and is altered to achieve reliability by integrating a technique presented in [2]. The integrated method should be able to achieve immediate authentication while preserving desired security and reliability and reducing memory requirements in sensor nodes.

An ad-hoc network based framework for monitoring brain function

Ad-hoc networks and mobile devices have become a crucial part of our daily lives. The low cost of wireless devices and free use of ad-hoc networks open an unlimited horizon to create new applications. Moreover, integrating several technologies can achieve almost unthinkable solutions. This paper presents a mobile solution framework to monitor human brain functions during real-life activities. The framework utilizes the internet, GSM wireless networks, Bluetooth technology and a number of data protocols, and consists of three main parts: a Bluetooth portable near-infrared light sensor; a personal digital assistant (PDA) and a personal computer (PC). The real-time data acquisition is performed by the sensor while mobility is provided by the GSM PDA. The data is sent over a various-protocol stack until it reaches the final destination (the host PC). The system provides a powerful light-weight human-brain-function monitoring system in real-life situations outside a lab environment. Several software components have been developed to achieve the integration of all these technologies and devices.