Milena Radenkovic

On the use of agents in a BioInformatics grid

My Grid is an e-Science Grid project that aims to help biologists and bioinformaticians to perform workflow-based in silico experiments, and help them to automate the management of such workflows through personalisation, notification of change and publication of experiments. In this paper, we describe the architecture of my Grid and how it will be used by the scientist. We then show how my Grid can benefit from agents technologies. We have identified three key uses of agent technologies in my Grid: user agents, able to customize and personalise data, agent communication languages offering a generic and portable communication medium, and negotiation allowing multiple distributed entities to reach service level agreements.

The multimedia challenges raised by pervasive games

Pervasive gaming is a new form of multimedia entertainment that extends the traditional computer gaming experience out into the real world. Through a combination of personal devices, positioning systems and other multimedia sensors, combined with wireless networking, a pervasive game can respond to a players movements and context and enable them to communicate with a game server and other players. We review recent examples of pervasive games in order to explain their distinctive characteristics as multimedia applications. We then consider the challenge of scaling pervasive games to include potentially very large numbers of players. We propose a new approach based upon a campaign model in which individuals, local groups and experts draw on a combination of pervasive games, online services and broadcasting to take part in national or even global events. We discuss the challenges that this raises for further research.

Reputation-based security protocol for MANETs in highly mobile disconnection-prone environments

This paper is concerned with fully distributed reputation-based mechanisms that improve security in MANETS. We introduce a number of optimisations to the current reputation schemes used in MANETs such as selective deviation tests and adaptive expiration timer that aim to deal with congestion and quick reputation convergence. We propose to use two different centrality measures for evaluation of the individual trust claims and resolving the aggregated ones. We design and build our prototype over AODV and test it in NS-2 in the presence of variable active blackhole attacks in highly mobile and sparse networks. Our results show that we achieve increased throughput while delay and jitter decrease and converge to AODV.

Multi-party distributed audio service with TCP fairness

Distributed Partial Mixing is an approach to creating a distributed audio service that supports optimisation of bandwidth utilization across multiple related audio streams (e.g. from concurrently active audio sources) while maintaining fairness to TCP traffic in best effort networks. Rate adaptation of streamed audio is difficult because of its rate sensitivity, the relatively limited range of encoding bandwidths available and the potential impact on the end user of rate-adaptation artefacts (such as changes of encoding). This paper describes and demonstrates how our design combines TCP-fairness with the stability that is desirable for streaming audio and other rate sensitive media. In particular, our design combines: a distributed multi-stream management/mixing architecture, loss event and round-trip time monitoring, rate limiting based on a TCP rate equation, tuned increase and decrease strategies and a loss-driven network probing mode. Experimental validation is performed over a wide range of network conditions including against various congesting levels, TCP and independent DPM traffic.

Efficient and adaptive congestion control for heterogeneous delay-tolerant networks

Detecting and dealing with congestion in delay-tolerant networks (DTNs) is an important and challenging problem. Current DTN forwarding algorithms typically direct traffic towards more central nodes in order to maximise delivery ratios and minimise delays, but as traffic demands increase these nodes may become saturated and unusable. We propose CafRep, an adaptive congestion aware protocol that detects and reacts to congested nodes and congested parts of the network by using implicit hybrid contact and resources congestion heuristics. CafRep exploits localised relative utility based approach to offload the traffic from more to less congested parts of the network, and to replicate at adaptively lower rate in different parts of the network with non-uniform congestion levels. We extensively evaluate our work against benchmark and competitive protocols across a range of metrics over three real connectivity and GPS traces such as Sassy 44], San Francisco Cabs 45] and Infocom 2006 33]. We show that CafRep performs well, independent of network connectivity and mobility patterns, and consistently outperforms the state-of-the-art DTN forwarding algorithms in the face of increasing rates of congestion. CafRep maintains higher availability and success ratios while keeping low delays, packet loss rates and delivery cost. We test CafRep in the presence of two application scenarios, with fixed rate traffic and with real world Facebook application traffic demands, showing that regardless of the type of traffic CafRep aims to deliver, it reduces congestion and improves forwarding performance.

Congestion aware forwarding in delay tolerant and social opportunistic networks

We propose an approach for opportunistic forwarding that supports optimization of multipoint high volume data flow transfer while maintaining high buffer availability and low delays. This paper explores a number of social, buffer and delay heuristics to offload the traffic from congested parts of the network and spread it over less congested parts of the network in order to keep low delays, high success ratios and high availability of nodes. We conduct an extensive set of experiments for assessing the performance of four newly proposed heuristics and compare them with Epidemic, Prophet, Spay and Wait and Spay and Focus protocols over real connectivity driven traces (RollerNet) and with a realistic publish subscribe filecasting application. We look into success ratio of answered queries, download times (delays) and availability of buffer across eight protocols for varying congestion levels in the face of increasing number of publishers and topic popularity. We show that all of our combined metrics perform better than Epidemic protocol, Prophet, Spray and Wait, Spray and Focus and our previous prototype across all the assessed criteria.

Promoting congestion control in opportunistic networks

This paper is concerned with congestion aware forwarding algorithms within opportunistic networks. We remove the reoccurring assumption of unlimited storage, and make it evident that congestion is a prominent problem that needs to be addressed. We propose a distributed congestion control algorithm that adaptively chooses the next hop based on contact history and statistics, as well as storage statistics. We aim to distribute the load away from the storage hotspots in order to spread the traffic around. We perform an extensive set of trace driven simulations for “several-to-many” communication patterns in opportunistic networks. We show that congestion control is an essential component in the transfer of data in opportunistic networks, and can be achieved in a fully open loop manner and by only local dissemination of statistics of nodes availability and connectivity. Our results with real connectivity traces show that by using novel availability heuristic we achieve higher levels of sent and delivered packets and outperform current opportunistic forwarding protocols, such as SimBetTS and FairRoute.

Practical MANETs for Pervasive Cattle Monitoring

The application of Mobile Ad Hoc Networks to cattle monitoring has the potential to increase the profitability of cattle production and positively impact the everyday live of farm personnel. The main research challenges are identifying and refining realistic requirements for a MANET routing protocol and designing such protocol. In this paper we report on the field experiments we performed in order to address this. Our approach builds on energy efficient MANETs to provide continuous monitoring of multidimensional parameters of animal mobility including temporal and spatial walking intensity and feed intake in order to detect oestrus, pregnancy, animal diseases and reduced efficiency of pastures. We support remote and in-situ, specific and range queries and notifications about newly detected events. Our extensive set of emulations show that we achieve lower and more balanced energy consumption while preserving the delays for increasing number of nodes within the user expected boundaries.

Deployment issues for multi-user audio support in CVEs

We describe an audio service for CVEs, designed to support many people speaking simultaneously and to operate across the Internet. Our service exploits a technique called Distributed Partial Mixing (DPM) to dynamically adapt to varying numbers of speakers and network congestion. Our DPM implementation dynamically manages the trade-off between congestion and audio quality when compared to the approaches of peer-to-peer forwarding and total mixing in a way that is fair to the TCP protocol and so operates as a “good Internet citizen”. This paper focuses on the large-scale deployment of DPM over wide area networks. In particular we raise and examine the issues when deploying DPM within the context of large dynamic environments. We argue that DPM paradigm remains feasible and desirable in such environments.

Framework for utility driven congestion control in delay tolerant opportunistic networks

Detecting and dealing with congestion in delay tolerant opportunistic networks is an important and challenging problem. In this paper we describe CAFREP, a unified congestion control framework for routing in such networks that adapts both data sending rates and data forwarding policies through a novel reactive fully distributed approach. CAFREP enables congestion control by detecting and reacting to congested nodes and congested parts of the network by using implicit hybrid contact and resources congestion heuristic. CAFREP exploits localized relative utility based approach to offload the traffic from more to less congested parts of the network, and replicate at adaptively lower rate in parts of the network with higher congestion. We extensively evaluate CAFREP against a number of state of the art adaptive and non-adaptive DTN routing protocols across a number of different metrics over three real trace driven experiments with different mobility and connectivity patterns such as Infocom 2006, Rollernet and Dieselnet CRAWDAD data sets. We show that CAFREP performs well in three different connectivity datasets and continuously outperforms four other protocols in terms of maintaining higher node availability and success ratio while keeping lower delays, lower packet loss rates and lower number of forwarded packets for increasing congestion levels.