Udayan Kumar

Integrating rapid risk mapping and mobile phone call record data for strategic malaria elimination planning

BackgroundAs successful malaria control programmes re-orientate towards elimination, the identification of transmission foci, targeting of attack measures to high-risk areas and management of importation risk become high priorities. When resources are limited and transmission is varying seasonally, approaches that can rapidly prioritize areas for surveillance and control can be valuable, and the most appropriate attack measure for a particular location is likely to differ depending on whether it exports or imports malaria infections.Methods/ResultsHere, using the example of Namibia, a method for targeting of interventions using surveillance data, satellite imagery, and mobile phone call records to support elimination planning is described. One year of aggregated movement patterns for over a million people across Namibia are analyzed, and linked with case-based risk maps built on satellite imagery. By combining case-data and movement, the way human population movements connect transmission risk areas is demonstrated. Communities that were strongly connected by relatively higher levels of movement were then identified, and net export and import of travellers and infection risks by region were quantified. These maps can aid the design of targeted interventions to maximally reduce the number of cases exported to other regions while employing appropriate interventions to manage risk in places that import them.ConclusionsThe approaches presented can be rapidly updated and used to identify where active surveillance for both local and imported cases should be increased, which regions would benefit from coordinating efforts, and how spatially progressive elimination plans can be designed. With improvements in surveillance systems linked to improved diagnosis of malaria, detailed satellite imagery being readily available and mobile phone usage data continually being collected by network providers, the potential exists to make operational use of such valuable, complimentary and contemporary datasets on an ongoing basis in infectious disease control and elimination.

PROTECT: proximity-based trust-advisor using encounters for mobile societies

Many interactions between network users rely on trust, which is becoming particularly important given the security breaches in the Internet today. These problems are further exacerbated by the dynamics in wireless mobile networks. In this paper we address the issue of trust advisory and establishment in mobile networks, with application to ad hoc networks, including DTNs. We utilize encounters in mobile societies in novel ways, noticing that mobility provides opportunities to build proximity, location and similarity based trust. Four new trust advisor filters are introduced - including encounter frequency, duration, behavior vectors and behavior matrices - and evaluated over an extensive set of real-world traces collected from a major university. Two sets of statistical analyses are performed; the first examines the underlying encounter relationships in mobile societies, and the second evaluates DTN routing in mobile peer-to-peer networks using trust and selfishness models. We find that for the analyzed trace, trust filters are stable in terms of growth with time (3 filters have close to 90% overlap of users over a period of 9 weeks) and the results produced by different filters are noticeably different. In our analysis for trust and selfishness model, our trust filters largely undo the effect of selfishness on the unreachability in a network. Thus improving the connectivity in a network with selfish nodes. We hope that our initial promising results open the door for further research on proximity-based trust.

On the efficacy of mobility modeling for DTN evaluation: Analysis of encounter statistics and spatio-temporal preferences

In mobile networking, the main goal of mobility modeling and simulation is the ability to accurately reproduce effects of realistic mobility on the performance of networking protocols. In the areas of adhoc and delay tolerant networks (DTNs), recent work on mobility modeling focused on replicating metrics of encounter statistics and spatio-temporal preferences. No studies have been conducted, however, to show whether matching these metrics is sufficient to accurately reproduce DTN protocol performance. In this study, we address this specific problem, and attempt to show the sufficiency (or lack thereof) of existing encounter and mobility metrics in reproducing realistic effects of mobility on networking protocols. We first analyze the characteristics of two well-established mobility models; the random direction and the time-variant community (TVC) models, and study whether they capture encounter statistics and preference patterns observed in real-world traces. Second, we contrast the performance of epidemic routing in DTNs based on the mobility models, to that based on extensive mobility traces. We provide two main findings. First, careful parameterization of the models can indeed replicate the metrics in question (e.g., inter-encounter time distribution). Second, even carefully crafted mobility models surprisingly result in protocol performance that is dramatically different from the trace-driven performance. The difference in message delivery delays can reach 67%, while difference in reachability approaches 80%. Such findings strongly suggest the need to revisit mobility modeling. Furthermore, they clearly show the insufficiency of existing encounter and preference metrics as a measure of mobility model goodness. Systematically establishing a new set of meaningful mobility metrics should certainly be addressed in future works.

Gender-based feature analysis in campus-wide WLANs

Rapid WLAN deployment has led to various research challenges and scenarios. In this paper we come up with a technique to classify users into social groups and then use this information to investigate the usage behavior of these groups. Grouping of users can be done based on diverse parameters like gender, major or other interest groups. In this paper we show the general methodology used to accomplish WLAN user groupings with an example of grouping by gender in a major university campus. The usage patterns of Males and Females are contrasted based on this grouping. Possible applications of this work encompass user profiling; studying gender gaps in social sciences; announcement/advertisement customization to target groups.

Extract: mining social features from WLAN traces--a gender-based case study

The next frontier for sensor networks is sensing the human society. Several mobile societies are emerging, especially with wide deployment of wireless LANs (WLANs) on campuses. WLAN traces can provide much insight into mobile user behavior. Such insight is essential to develop realistic models and to design better networks, and analyze effects of social attributes on mobile network usage. The most extensive libraries of wireless traces are collected from university campuses, are anonymized and do not provide affiliation, gender or preference information explicitly. Hence, it becomes a challenge to analyze network usage characteristics for social groups using the existing traces. In this paper, we present two novel scientific techniques to classify WLAN users into social groups. The first technique uses mapping of the traces into buildings (e.g., dept. buildings, libraries, sororities and fraternities) to extract affiliation and gender information based on network usage statistics. The second technique utilizes directory information that can be linked to WLAN users to extract useful information. For example, usernames of the WLAN users (if available) can be used to find users gender based on first name and databases. As a case study we perform classification and behavior analysis of users by gender. Extensive WLAN traces from two major universities are collected over three years and analyzed. Results from both the methods are cross-validated and show more than 90% correspondence. Results of gender classification are then used to examine usage patterns and preferences across gender groups, including spatio-temporal distribution of wireless on-line activity, study majors and vendor preference. In some cases these metrics are equal across genders, however, there are several interesting cases that clearly indicate statistically significant and consistent effects of gender; e.g., males have longer on-line sessions in Engineering and Music, while females have longer sessions in Social Sciences and Sports areas. At one university female groups consistently preferred Apple computers. These findings can have a great impact on several mobile networking applications; they can be directly used for realistic modeling of wireless user on-line behavior, mobility and virus susceptibility, and for designing socially-aware protocols and class-based or gender-based services, to name a few.

Visualization and Representation of Mobile Network Users

In this paper we discuss a framework to visualize user dynamics in large-scale WLAN traces on the Google Earth interface. The visualization technique helps us to observe the users within the temporal and geographic context, and it is a powerful tool to show case several earlier research results, including user classification and behavior-aware protocol design. We plan to use the visualization tool to help user mobility/encounter pattern prediction in the future.

Proximity based trust-advisor using encounters for mobile societies: Analysis of four filters

Many interactions between network users rely on trust, which is becoming particularly important given the security breaches in the Internet today. These problems are further exacerbated by the dynamics in wireless mobile networks. In this paper, we address the issue of trust advisory and its establishment in mobile networks, with application to ad hoc networks, including delay tolerant (DTNs). We utilize encounters in mobile societies in novel ways, noticing that mobility provides opportunities to build proximity, location, and similarity based trust. Four new trust advisor filters are introduced – including encounter frequency, duration, behavior vectors, and behavior matrices. The filters are evaluated over an extensive set of real-world traces collected from a major university. Two sets of statistical analyses are performed; the first examines the underlying encounter relationships in mobile societies, and the second evaluates DTN routing in mobile peer-to-peer networks using trust and selfishness models. We find that for the analyzed trace, trust filters are stable in terms of growth with time (three filters have close to 90% overlap of users over a period of 9 weeks) and the results produced by different filters are noticeably different. In our analysis for trust and selfishness model, our trust filters largely undo the effect of selfishness on the unreachability in a network. Thus improving the connectivity in a network with selfish nodes. We hope that our initial promising results open the door for further research on proximity-based trust. Copyright

Polite: A policy framework for building managed mobile apps

The proliferation of smart phones inside enterprises and the number of enterprise apps (applications) available for various smart phone platforms has been increasing. This trend is expected to continue as smart phones tend to become the device of choice to access both enterprise and personal data. Making enterprise sensitive data accessible on smart phones requires that adequate protection mechanisms be available on these devices to ensure that sensitive data is not compromised due to various reasons, such as employees losing phones to malicious apps (installed by the user) running on the phones. Most of the existing solutions either provide device level control or have an external agent monitoring the applications behavior, and has numerous limitations. In this paper we propose a framework, Polite, to build enterprise mobile apps that can be managed at run-time, which is less intrusive to the end user while providing stronger security guarantees to the enterprise. We describe several critical scenarios where controlling the run time behavior of apps on the phone is essential and how our architecture can provide security guarantees that are not possible with existing solutions. Performance results of our implementation indicate that our framework induces a minimal overhead of only 6% that may be acceptable for most enterprise mobile apps.

Proximity-based trust-advisor using encounters

No trust tantamounts to no communication in DTN and mobile ad-Hoc networks. In this work we propose a novel encounter-based trust framework based on the principle of homophiliy. We propose 4 trust advisors and analyze their performance using WLAN traces. We find that our filters not only provide meaningful trust but also improve network performance in the presence of selfish nodes.

Discovering trustworthy social spaces

Many future mobile services and applications will center on the social and community aspects of mobile societies. Interactions and connections between users in mobile networks are usually subject to the strength of the connections between the nodes, informed by historical events. This study proposes, implements and evaluates novel methods to dynamically measure the strength of social connections and similarity based on historical mobility behavior and encounter information. Through our protocol and application we investigate the feasibility of discovering known encountered devices, in addition to the opportunistic identification of potentially-strong new connections. We propose a set of 4 filters to rate and rank mobile encounters identifying users with similar behavior. We have developed and deployed ConnectEnc application on Android and Nokia N810 platform to measure the link between the scores of proposed filters and the existence (or lack) of social relationship with the rated devices. We find that a statistically strong relationship exists between our recommendation and social relationship with the devices rated by the users (for LVC, r=0.84, p <0.01). With this similarity based trustworthy node discovery, several potential applications can be enabled including mobile social networking, building groups and communities of interest, localized alert and emergency notification, context-aware and similarity-based networking.