Christopher B. Mayer

Optimization of media processing workflows with adaptive operator behaviors

In this paper, we present the ARIA media processing workflow architecture that processes, filters, and fuses sensory inputs and actuates responses in real-time. The components of the architecture are programmable and adaptable; i.e. the delay, size, and quality/precision characteristics of the individual operators can be controlled via a number of parameters. Each data object processed by qStream components is subject to transformations based on the parameter values. For instance, the quality of an output data object and the corresponding processing delay and resource usage depend on the values assigned to parameters of the operators in the object flow path. In Candan, Peng, Ryu, Chatha, Mayer (Efficient stream routing in quality- and resource-adaptive flow architectures. In: Workshop on multimedia information systems, 2004), we introduced a class of flow optimization problems that promote creation and delivery of small delay or small resource-usage objects to the actuators in single-sensor, single-actuator workflows. In this paper, we extend our attention to multi-sensor media processing workflow scenarios. The algorithms we present take into account the implicit dependencies between various system parameters, such as resource consumption and object sizes. We experimentally show the effectiveness and efficiency of the algorithms.

The effectiveness of dynamic ant colony tuning

We examine the Genetically Modified Ant Colony System (GMACS) algorithm [3], which claims to dynamically tune an Ant Colony Optimization (ACO) algorithm to its near-optimal parameters. While our research indicates that the use of GMACS does result in higher quality solutions over a hand-tuned ACO algorithm, we found that the algorithm is ultimately hindered by its emphasis on randomized ant breeding. Specifically, our investigation shows that tuning ACO parameters on a single colony using a genetic algorithm, as done by GMACS, is not as effective as it may first appear and has several drawbacks.

Enabling Skip Graphs to Process K-Dimensional Range Queries in a Mobile Sensor Network

A skip graph is a resilient application-layer routing structure that supports range queries of distributed k-dimensional data. By sorting deterministic keys into groups based on locally computed random membership vectors, nodes in a standard skip graph can optimize range query performance in mobile networks such as unmanned aerial vehicle swarms. We propose a skip graph extension that inverts the key and membership vector roles and bases group membership on deterministic vectors derived from the z-ordering of k-dimensional data and sorting within groups is based on locally computed random keys.

Fairly Redistributing Failed Server Load in a Distributed System

We recently proposed a novel method for large-object replication and load balancing. Our method is particularly well-suited to data grids, data warehousing providers, and hosting of dynamic web sites. The method attempts to distribute object request load fairly to servers according to server capacity so that the likelihood of them overloading, and hence failing, is reduced. Unfortunately, server failures cannot be eliminated entirely. When a server fails, the load carried by that server must be absorbed by the rest of the system. Unless this load is distributed fairly across the remaining servers, they may also overload, creating a cascade of failures and reduced quality of service. In this paper, we propose an efficient method for fairly redistributing the load of a failed server or set of failed servers within our replication system. We also report on experimental results that verify the validity of our approach.

Assessing Standard and Inverted Skip Graphs Using Multi-Dimensional Range Queries and Mobile Nodes

The skip graph, an application-layer data structure for routing and indexing, may be used in a sensor network to facilitate queries of the distributed k-dimensional data collected by the nodes, such as their geographic positions. Nodes in a standard skip graph sort keys in layered groups, where group membership is determined by random membership vectors. We propose a skip graph extension that inverts the key and membership vector roles, wherein group membership is based on deterministic z-ordering of k-dimensional data and sorting within groups is based on random keys. This extension retains the structure of a standard skip graph, and can be modified for use in 3D environments such as unmanned aerial vehicle swarms. Our results indicate this extension reduces the number of nodes contacted relative to the query precision, the volume of the multidimensional space, and the skip graphs height. We discuss our extensions performance relative to metrics at the application and network layers for static and mobile 2D networks.

Replicating multi-quality web applications using ACO and bipartite graphs

This paper presents the application of the Ant Colony Optimization (ACO) meta-heuristic to a new NP-hard problem involving the replication of multi-quality database-driven web applications (DA s) by a large application service provider (ASP). The ASP must assign DA replicas to its network of heterogeneous servers so that user demand is satisfied at the desired quality level and replica update loads are minimized. Our ACO algorithm, AntDA , for solving the ASP’s replication problem has several novel or infrequently seen features: ants traverse a bipartite graph in both directions as they construct solutions, pheromone is used for traversing from one side of the bipartite graph to the other and back again, heuristic edge values change as ants construct solutions, and ants may sometimes produce infeasible solutions. Testing shows that the best results are achieved by using pheromone and heuristics to traverse the bipartite graph in both directions. Additionally, experiments show that AntDA outperforms several other solution methods.