Jong Kook Kim

An overview of MSHN: the Management System for Heterogeneous Networks

The Management System for Heterogeneous Networks (MSHN) is a resource management system for use in heterogeneous environments. This paper describes the goals of MSHN, its architecture, and both completed and ongoing research experiments. MSHNs main goal is to determine the best way to support the execution of many different applications, each with its own quality of service (QoS) requirements, in a distributed, heterogeneous environment. MSHNs architecture consists of seven distributed, potentially replicated components that communicate with one another using CORBA (Common Object Request Broker Architecture). MSHNs experimental investigations include: the accurate, transparent determination of the end-to-end status of resources; the identification of optimization criteria and how non-determinism and the granularity of models affect the performance of various scheduling heuristics that optimize those criteria; the determination of how security should be incorporated between components as well as how to account for security as a QoS attribute; and the identification of problems inherent in application and system characterization.

Dynamic mapping in a heterogeneous environment with tasks having priorities and multiple deadlines

In a distributed heterogeneous computing system, the resources have different capabilities and tasks have different requirements. To maximize the performance of the system, it is essential to assign resources to tasks (match) and order the execution of tasks on each resource (schedule in a manner that exploits the heterogeneity of the resources and tasks. The mapping (defined as matching and scheduling) of tasks onto machines with varied computational capabilities has been shown, in general, to be an NP-complete problem. Therefore, heuristic techniques to find a near-optimal solution to this mapping problem are required. Dynamic mapping is performed when the arrival of tasks is not known a priori. In the heterogeneous environment considered in this study, tasks arrive randomly, tasks are independent (i.e., no communication among tasks), and tasks have priorities and multiple deadlines. This research proposes, evaluates, and compares eight dynamic heuristics. The performance of the best heuristics is 83% of an upper bound.

A metric and mixed-integer-programming-based approach for resource allocation in dynamic real-time systems

Dynamic real-time systems such as embedded systems operate in environments in which several parameters vary at run time. These systems must satisfy several performance requirements. Resource allocation on these systems becomes challenging because variations of run-time parameters may cause violations of the performance requirements. Performance violations result in the need for dynamic re-allocation, which is a costly operation. A method for allocating resources such that the allocation can sustain the system in the light of a continuously changing environment is developed. We introduce a novel performance metric called MAIL (maximum allowable increase in load) to capture the effectiveness of a resource allocation. Given a resource allocation, MAIL quantifies the amount of additional load that can be sustained by the system without any performance violations. A mixed-integer-programming-based approach (MIP) is developed to determine a resource allocation that has the highest MAIL value. Using simulations, several sets of experiments are conducted to evaluate our heuristics in various scenarios of machine and task heterogeneities. The performance of MIP is compared with three other heuristics: integer-programming based, greedy, and classic min-min. Our results show that MIP performs significantly better when compared with the other heuristics.

Dynamic mapping in energy constrained heterogeneous computing systems

An ad hoc grid is a wireless heterogeneous computing environment without a fixed infrastructure. The wireless devices have different capabilities, have limited battery capacity, support dynamic voltage scaling, and are expected to be used for eight hours at a time and then recharged. To maximize the performance of the system, it is essential to assign resources to tasks (match) and order the execution of tasks on each resource (schedule) in a manner that exploits the heterogeneity of the resources and tasks while considering the energy constraints of the devices. In the single-hop ad hoc grid heterogeneous environment considered in this study, tasks arrive unpredictably, are independent (i.e., no precedent constraints for tasks), and have priorities and deadlines. The problem is to map (match and schedule) tasks onto devices such that the number of highest priority tasks completed by their deadlines during eight hours is maximized while efficiently utilizing the overall system energy. A model for dynamically mapping tasks onto wireless devices is introduced. Seven dynamic mapping heuristics for this environment are designed and compared to each other and to a mathematical bound.

Using DVFS and task scheduling algorithms for a hard real-time heterogeneous multicore processor environment

The usage of heterogeneous multicore processors (HMP) are rapidly spreading from data centers for large scale deployment to smart phones for the flexibility to adapt to power constraints and performance needs. In this paper, we show that for an example HMP environment, an intelligent-task scheduler is critical in improving performance and energy efficiency. The environment in this paper assumes that the tasks are independent, have hard real-time constraints, and a multicore systems where processors can be manipulated to change the clock cycle speed and power levels. Tasks are assumed to arrive aperiodically and these tasks are applications from the SPEC CPU 2006 benchmark suite. For evaluation, an actual system composed of two multicore processors which support on-the-fly DVFS is used in this study. One of our energy efficient algorithms achieved 49.9% higher task completion rate than an enhanced version of the naïve Linux scheduler while consuming only 45.3% of the energy.

Applying passive RFID system to wireless headphones for extreme low power consumption

One of major design concerns about wireless headphone is power consumption. In this paper, we propose a novel design for extreme low power headphone implementation by extending the EPC Class-1 Generation2 RFID protocol for delivering stream data. We prototyped a reader as a stream generator, and a passive tag as an audio receiver to consume less power than any other protocols for wireless headphones.

Enhancing the performance of a distributed mobile computing environment by topology construction

Topology construction methods for a distributed mobile computing environment where the devices are heterogeneous, mobile, and use dynamic voltage scaling and variable transmission power control methods to efficiently use the overall system energy are developed in this research. The final goal of the research is to complete as many tasks as possible using the distributed mobile computing system. The tasks in this system are heterogeneous and must be completed by their deadline to have value. The tasks must be intelligently distributed among the devices to efficiently use the system resources. The reason for the new topology methods was that as the number of devices increase for the example environment the number of communications dropped because of communication collision increased. We propose two major ideas for topology algorithms to enhance the performance and compared it with the all-connected environment and the one that uses the MSMR method which showed improved performance over previous ones. Different methods proved to be better than MSMR in different scenarios for the distributed mobile computing environment.

Semantic hashtag relation classification using co-occurrence word information

Social Networking Service users express their thoughts and feelings using hashtags. Hashtags can be related to other hashtags and these hashtags and images are used together in a post that the user wrote. Understanding the meaning of a hashtag is one of the ways to learn latent semantic expressions of words. Existing methods for learning semantic analysis use large corpus. This research focuses on the classification of semantic words using a users hashtag data and co-occurrence hashtag information.

Energy-Efficient Stabilized Automatic Control for Multicore Baseband in Millimeter-Wave Systems

The fifth generation (5G) cellular network is upon us. Academia and Industry have intensively collaborated together to bring the power of 5G cellular networks to the masses, and now the 5G millimeter-wave (mmWave) platforms come into being in the market. One of the most popular 5GmmWave platforms mounts the massive mmWave phased antenna arrays in order to transfer a huge number of bits in a second (e.g., more than ten gigabits-per-second) to the baseband in the platform. While exploiting chip multicore processors (CMPs) may be the best solution to process such huge data in the mmWave baseband platform, power dissipate by the CMPs should become critical. Starting from an intuition that utilizing all processors in every single time introduces inefficient energy consumption, this paper proposes an energy-aware queue-stable control (EQC) algorithm to control the activation/deactivation of individual processors and antenna arrays for pursuing time average energy consumption minimization subject to the stability of queues in the 5G-mmWave baseband. Results from intensive simulations based on realistic experimental setups demonstrate the efficacy of the proposed EQC that achieves significant energy savings while queue stability is maintained.

Information propagation modeling in a drone network using disease epidemic models

Drones transmit and receive packets with each other in a wireless network setting. Packet transmission among drones fails for various reasons. The pattern of information propagation through the packet transmission in a drone network can be considered similar to the pattern of infectious disease transmission in a human interaction network. In this work, we use a Microscopic Markov Chain Approach (MMCA), which has been applied to model the patterns of disease epidemics in a human network, to investigate the packet transmission pattern in a microscopic scale. Throughout the simulation studies, we investigated the usefulness of MMCAs for a drone network.