H. M. N. Dilum Bandara

Collaborative applications over peer-to-peer systems–challenges and solutions

Emerging collaborative Peer-to-Peer (P2P) systems require discovery and utilization of diverse, multi-attribute, distributed, and dynamic groups of resources to achieve greater tasks beyond conventional file and processor cycle sharing. Collaborations involving application specific resources and dynamic quality of service goals are stressing current P2P architectures. Salient features and desirable characteristics of collaborative P2P systems are highlighted. Resource advertising, selecting, matching, and binding, the critical phases in these systems, and their associated challenges are reviewed using examples from distributed collaborative adaptive sensing systems, cloud computing, and mobile social networks. State-of-the-art resource discovery/aggregation solutions are compared with respect to their architecture, lookup overhead, load balancing, etc., to determine their ability to meet the goals and challenges of each critical phase. Incentives, trust, privacy, and security issues are also discussed, as they will ultimately determine the success of a collaborative P2P system. Open issues and research opportunities that are essential to achieve the true potential of collaborative P2P systems are discussed.

A Top-Down Clustering and Cluster-Tree-Based Routing Scheme for Wireless Sensor Networks

Cluster-based organization of large sensor networks is the basis for many techniques aimed at enhancing power conservation and network management. A backbone network in the form of a cluster tree further enhances routing, broadcasting, and in-network processing. We propose a configurable top-down cluster and cluster-tree formation algorithm, a cluster-tree self-optimization phase, a hierarchical cluster addressing scheme, and a routing scheme. Such self-organization makes it possible to effectively deliver messages to a sink as well as within the network. For example, a circular sensor field with a sink in the center can establish cross-links and circular-paths within the cluster tree to deliver messages through shorter routes while reducing hotspots and consequently increasing network lifetime. Cluster and cluster-tree formation algorithm is independent of physical topology, and does not require a priori neighborhood information, location awareness, or time synchronization. Algorithm parameters allow for formation of cluster trees with desirable properties such as controlled breadth/depth, uniform cluster size, and circular clusters. Characteristics of clusters, cluster tree, and routing are used to demonstrate the effectiveness of the schemes over existing techniques.

An enhanced top-down cluster and cluster tree formation algorithm for Wireless Sensor Networks

Clustering is a key technique to simplify network management while enabling power conservation and reduced channel contention in large scale Wireless Sensor Networks. A hierarchy of clusters in the form of a cluster tree can further enhance upper layer functions such as routing, broadcasting and query delivery. We propose a generic top-down cluster and cluster tree formation algorithm that does not depends on neighborhood information, location awareness, time synchronization and network topology. It also scales well into large networks. By varying parameters in the algorithm, cluster tress with desirable properties such as controlled breadth and depth, uniform cluster size and improved circularity can be achieved. Different characteristics of clusters and cluster trees are evaluated using simulation based results.

A peer-to-peer collaboration framework for multi-sensor data fusion

A peer-to-peer collaboration framework for multi-sensor data fusion in resource-rich radar networks is presented. In this high data volume real-time application, data from multiple radars are combined to improve the accuracy of radar scans (e.g., correct for attenuation) and to provide a composite view of the area covered by the radars. Data fusion process is subject to two constraints: (1) the accuracy requirement of the final fused results, which may be different at different end nodes, and (2) the real-time requirements of the application. The accuracy requirement is achieved by dynamically selecting the appropriate set of data to exchange among the multiple radar nodes. A mechanism for selecting a dataset based on current application-specific needs is presented. We also present a dynamic peer-selection algorithm, Best Peer Selection (BPS), that chooses a set of peers based on their computation and communication capabilities to minimize the data processing time per integration algorithm. Simulation-based results show that BPS can deliver a significant performance improvement, even when the peers have high variability in available network and computation resources.

Evaluation of P2P resource discovery architectures using real-life multi-attribute resource and query characteristics

Emerging collaborative Peer-to-Peer (P2P) applications rely on resource discovery solutions to aggregate groups of heterogeneous, multi-attribute, and dynamic resources that are distributed. In the absence of data and understanding of real-life resource and query characteristics, design and evaluation of existing solutions have relied on many simplifying assumptions. We first present a summary of resource and query characteristics from PlanetLab. These characteristics are then used to evaluate fundamental design choices for multi-attribute resource discovery based on the cost of advertising/querying resources, index size, and load balancing. Simulation-based analysis indicates that the cost of advertising dynamic attributes is significant and in-creases with the number of attributes. Compared to uniform queries, real-world queries are relatively easier to resolve using unstructured, superpeer, and single-attribute dominated query based structured P2P solutions. However, they cause significant load balancing issues in all the designs where a few nodes are mainly involved in answering majority of queries and/or indexing resources. Moreover, cost of resource discovery in structured P2P systems is effectively O(N) as most range queries are less specific. Thus, many existing design choices are applicable only under specific conditions and their performances tend to degrade under realistic workloads.

On Characteristics and Modeling of P2P Resources with Correlated Static and Dynamic Attributes

Modeling and simulation of Peer-to-Peer (P2P) resources with correlated static and dynamic attributes is essential in application design, validation, and performance analysis. A novel mechanism is presented to generate realistic synthetic traces of multivariate static and dynamic attributes of P2P resources. The methodology is demonstrated using characteristics of PlanetLab node traces. First, a multi-attribute resource model is defined using a selected set of static and dynamic attributes. Second, characteristics of resources are presented. We observe that attribute values are correlated, follow a mixture of probability distributions, and time series of some of the dynamic attributes are nonstationary. Third, random vectors of static attributes are generated using empirical copulas that capture the entire dependence structure of multivariate distribution of attributes. Finally, time series of dynamic attributes are randomly drawn from a library of multivariate-time-series segments extracted from PlanetLab traces. These segments are identified by detecting the structural changes in time series corresponding to a selected attribute. Time series corresponding to rest of the attributes are split at the same breakpoints and randomly drawn together to preserve their contemporaneous correlation. Furthermore, a tool is developed to automate the synthetic data generation process and its output is validated using statistical tests.

Characteristics of multi-attribute resources/queries and implications on P2P resource discovery

Though resource discovery is a fundamental requirement in collaborative peer-to-peer, grid, and cloud computing, very little is known about resource/query characteristics and their impact on resource discovery. Fundamental design choices for distributed resource advertising and querying are evaluated in the context of existing practical systems. First, a generic model for cost of resource discovery is presented. Second, multi-attribute resource and query characteristics from Planet-Lab and SETI@home are presented. We observe that attributes of both resources and queries are highly skewed, correlated, queries are less specific, and Generalized Pareto distribution is suitable for capturing the distribution of most dynamic attributes and their rate of change. Based on these observations, different design choices are evaluated for resource discovery in terms of their cost of advertising/querying, latency, load balancing, and routing table size. The findings indicate that superpeer-based architectures have the potential to support large-scale resource aggregation as they simultaneously balance the cost and load.

Distributed, multi-user, multi-application, and multi-sensor data fusion over named data networks

Named Data Networking (NDN) routes data based on their application-layer content names enabling location independence, in-network caching, and enhanced security. A proof-of-concept solution is presented that demonstrates the applicability of NDN for multi-user, multi-application, and multi-sensor data-fusion systems. The system consists of a collaborative network of weather radars name data based on their geographic location and weather feature (e.g., reflectivity of clouds and wind velocity). This enables end users to specify an area of interest for a particular weather feature while being oblivious to the placement of radars and associated computing facilities. Conversely, the data-fusion system can also use its knowledge about the underlying system to decide the best sensing and data processing strategies. Such sensor-independent names also enhance resilience, enable processing data close to the source, and benefit from NDN features such as in-network caching and duplicate query suppression, consequently reducing the bandwidth requirements of the entire data-fusion system. The solution is implemented as an overlaid NDN enabling the benefits of both the NDN and overlay networks. Simulation-based analysis using reflectivity data from an actual weather event showed 84% reduction in peak bandwidth consumption of radars and 95% reduction in peak query resolution latency.

Exploiting Communities for Enhancing Lookup Performance in Structured P2P Systems

Large Peer-to-Peer (P2P) systems for file transfer exhibit the presence of communities based on semantic, geographic, or organizational interests of users. Generally, resources commonly shared within individual communities are relatively unpopular and inconspicuous in the system-wide behavior. These communities are unable to benefit significantly from performance enhancement schemes such as caching that focus only on the most dominant queries. We propose a generic caching framework that enhances lookup performance of individual communities while providing even better performance to the dominant communities. The caching framework can be used with any structured P2P system that provides alternative paths to a given destination. Furthermore, the solution is adaptive to changing popularity and user interests, works with any skewed distribution of queries, needs small caches, utilizes local statistics, and introduces minimal modifications and overhead to the overlay network. Simulations based on Chord overlay show 40% reduction in average path length with individual communities indicating three times improvement in performance over system-wide caching.

Workload and Resource Aware Proactive Auto-scaler for PaaS Cloud

Elasticity is a key feature in Cloud Computing where virtualized resources are provisioned and de-provisioned via auto-scaling. However, auto-scaling in most Platform-as-a-Service (PaaS) systems is based on reactive, threshold-driven approaches. Such systems are incapable of catering to rapidly varying workloads, unless the associated thresholds are sufficiently low. Alternatively, maintaining low thresholds leads to resource over-provisioning under relatively stable workloads. Moreover, thresholds are not a good indication of QoS compliance, which is a key performance indicator of a cloud application. Hence, it is nontrivial to determine an optimum threshold while minimizing costs and meeting QoS demands. We propose inteliScaler, a proactive and cost-aware auto-scaling solution to address these issues by combining a predictive model, cost model, and a smart killing feature. An ensemble workload prediction mechanism is introduced based on time series and machine learning techniques for making accurate predictions on drastically different workload patterns. Utility of the solution is demonstrated using both simulations and empirical evaluations using Apache Stratos PaaS (deployed on the AWS EC2), as well as RUBiS and real-world workload traces. Results show significant QoS improvements and cost reductions by inteliScaler compared to a typical reactive and threshold-based PaaS auto-scaling solution.