Ajantha S. Atukorale

Bassa: a time shifted web caching system for developing regions

This paper presents a time shifted web caching system named Bassa. It is capable of scheduling, fetching and caching of web objects larger than a given size threshold to take place within the off peak time of the network. We have achieved promising results using this system by maximizing the use of available bandwidth for more interactive web objects during peak hours while utilizing previously underutilized band-width in off peak hours. By empirically analyzing the effect of time shifted caching on users we have unearthed that such systems encourage them to download larger files more than they did previously while accessing websites with more multimedia content such as falsh videos during the peak hours. In this paper we present those findings empirically based on data gathered from proxy server logs at University of Colombo School of Computing.

ANDROPHSY - forensic framework for Android

There have been many studies addressed towards various sections in the arena of Android platform. However conducting a digital forensics investigation using Open Source tools in Android platform is still a challenging task. As of the date of this study, there is no standard Open Source or noncommercial tool available to support the entire life cycle of the digital forensics investigation methodology. In order to bridge this gap in open source forensics tool for Android platform, the work implemented a user friendly, feature rich, open source, mobile forensic framework for Android platform that supports all phases of a full digital forensic investigation process.

A robust algorithm for determining the newsworthiness of microblogs

Microblogging platforms such as Twitter have become a primary medium for people to share their experiences and opinions on a broad range of topics. Because posts on Twitter are publicly viewable by default, Twitter can be used to get up-to-date information on events like natural disasters, disease outbreaks or sports events. Building a cohesive summary out of tweets on long running events is an interesting problem which research community is interested in. But the abundance of tweets containing user opinions and their sentiments towards a topic necessitates the need of extracting newsworthy tweets from a large stream of tweets on a single topic. But most of such methods require large hand-labeled corpora to be used for training the model. But this is not practical for a rapidly updating medium like Twitter. In this paper we address this problem with the introduction of a novel heuristic based annotation scheme to generate training dataset for the system. A hand-labeled corpus of tweets is only used for benchmarking the objectivity classifier. Our classifier could achieve an F1-score of 80% on a manually annotated gold standard dataset.

The Relationship Between Psychological Distress and Human Computer Interaction Parameters: Linear or Non-linear?

During a survey based study, it has been revealed that psychological distress is reflected in the way computer users interact with computers. Based on the feedback of the subjects, the types of stress-reflector interactions were also identified. However, the specific interactions related to the stress, the nature of the relationship between the identified ‘stress-reflectors’ and the actual level of psychological distress was yet to be investigated. Accordingly, the initial survey was then followed by a series of activities with the objective of uncovering the nature of the relationship exists between the two. Initially, the most frequented interaction behaviors of computer users under stress were identified. Secondly these behaviors were recorded using a background analyzer together with a simultaneous measure of the distress level of the relevant computer user. Finally, behavior measures of individuals were mapped with their personal distress scores, producing the dataset using which the nature of correlations were explored. This paper presents the work carried out during this exploration, keeping the initial focus on linear methods, and provides justifications for the claim: The relationship between the distress score and the human computer interaction parameters is non-linear.

Decentralized reputation based trust model for peer-to-peer content distribution networks

Enormous content distribution systems improved broadly with the rapid growth of novel and innovative technologies. Peer-to-peer (P2P) content distribution network (CDN) technologies such innovative technological improvement which claims low cost, efficient high demand data distribution and it gradually involves to the next-generation CDNs. Communications among the nodes of such open network infrastructures commonly perceived as an environment offering both opportunities and threats, which are based on trust issues. Thus, a trust mechanism is required to establish secure communication between the content delivery nodes. This paper introduces a novel, decentralized, cooperative and self-organized reputation based trust algorithm to mitigate the security complications of CDNs. To illustrate the trust model, a novel P2P hybrid network infrastructure has been introduced. The key notion of the research is to validate trust of the target before send or accept the traffic. Furthermore, performance of the proposed trust algorithm is evaluated through the ns-2 simulator in order to analyze the trusted and untrusted behavior of the P2P content delivery nodes.

Network data classification using graph partition

Application of network classification can be seen in many domains. These varies from preserving the quality of network to analyzing personal characteristics of network users. However current methods applied for network data classification does not meet the expectations. This is because networks are dynamic which are prone to rapid changes, while methods used for the classification has been either trained using examples or defined using heuristics. World Wide Web itself is a big graph which is made out of number of URLS connecting each other via hyper-links. Hence in this work we have used this graph nature of WWW and applied graph theories to partition the network to classify network data. We have used results obtained by classifying the network traffic using k-means algorithm to evaluate the performance and usability of proposed method.

A novel classifier for engineering web traffic

This work presents a classifier capable of classifying websites accessed by a community of users based on characteristics of web traffic they generate. The classifier introduced here can be used for engineering web traffic. This classifier is capable of identifying a few different classes of websites based on intensity of bursts of web requests made to each website in both the long and the short run. Previous studies have proposed different classification algorithms in the area of general network traffic classification and web traffic mining. Work presented in this paper is based on Kohonens self organizing map trained using web traffic traces gathered from a busy proxy server. By integrating this classifier with a rate controlling module of a proxy server, web traffic can be engineered very easily.