Daeyeon Park

Railroad: virtual infrastructure for data dissemination in wireless sensor networks

In this paper, we present Railroad, a data dissemination architecture for large-scale wireless sensor networks. Railroad system proactively exploits a virtual infrastructure called Rail, which is an area where all the metadata of event data are stored. There is only one Rail in the network and it acts as a rendezvous area of the events and the queries. Rail is placed in the middle area of the field so that every node can easily access it. Once a query is issued, it circulates around Rail and searches relevant data stored in Rail. When a relevant metadata is found, the source node of the data transmits the corresponding data to the sink node which has issued the query. By using Rail, Railroad achieves a scalable and energy-efficient data dissemination architecture under dynamic conditions with multiple mobile observers and targets. We evaluate and compare the communication cost and the hot spot message complexity of Railroad with previous approaches.

Adaptive core multicast routing protocol

The Adaptive Core Multicast Routing Protocol (ACMRP) is proposed for multicast routing in ad hoc networks. ACMRP is on demand core-based multicast routing protocol that is based on a multicast mesh. In ACMRP, a core is not well-known and it adapts to the current network topology and group membership. The enhanced adaptivity minimizes the core dependency and, accordingly, improves performance and robustness of ACMRP. A multicast mesh is created and maintained by the periodic flooding of the adaptive core. Since the flooding traffic is evenly maintained and a mesh provides rich connectivity among group members, ACMRP can achieve efficiency, scalability, and effectiveness. We evaluate scalability and performance of ACMRP via simulation.

A virtual infrastructure for large-scale wireless sensor networks

The primary goal of a wireless sensor network is to collect useful information from the network. Most wireless sensor networks are assumed that the number of nodes are very large and they should operate with confined resources. Consequently it is important to take a scalable and energy-efficient architecture. In this paper, we present Railroad, a data collection and topology management architecture for large-scale wireless sensor networks. It proactively exploits a virtual infrastructure called Rail, which acts as a rendezvous area of the event data and queries. By using Rail, Railroad achieves scalability and energy efficiency under dynamic conditions with multiple mobile observers and targets. We evaluate the communication cost and the hot area message complexity of Railroad and compare them with previous approaches. We evaluate communication cost of Railroad by both an analytic model and simulations.

A lightweight personal grid using a supernode network

Although grid computing has become popular in scientific research and the enterprise domain, there is no way a personal user who does not belong to any organization can construct or access the grid system for solving her problems. The main goal is to propose a lightweight personal grid that is self-organized without complicated configuration and infrastructural support like central servers and self-managed with the cooperation of participants. The key component of our system is the supernode network designed for autonomously and efficiently constructing the grid system. Based on the supernode network, our personal grid system forms a virtual organization (VO) with nodes having physical proximity. Therefore, communication latency between nodes in a VO can be minimized. Moreover, hierarchical scheduling is adopted using the two-level structure of the supernode network and new service architecture is proposed to improve the reusability of applications.

Request rate adaptive dispatching architecture for scalable Internet server

A scalable web server is one of popular information replication methods to accomodate the rapidly growing World Wide Web. It is basically a server clustering technique which supplies information from several servers with the illusion of one serving entity. Request distribution among the back-end servers is crucial work for the scalable web server performance. One popular approach for request distribution is using centralized scheduler such as dispatcher. It provides full control of scheduling and fine-grained load balancing but it suffers from potential bottleneck. DNS based approach utilizes name resolution process for request distribution. It is simple and low overhead scheduling method but it has limited control of load balancing because of the name resolution caching in the intermediate name servers. Another approach such as server based scheme has released the bottleneck problem by distributing the scheduling ability to all participating servers. However, the dispatching efficiency degrades because the redirection process shares resources with data processing jobs. We have developed an adaptive load balancing method that changes the number of scheduling entities according to different workload. It behaves exactly like dispatcher based scheme with low or intermediate workload, taking advantage of fine-grained load balancing. When the dispatcher is overloaded, the dispatching job is partially distributed to other entities such as DNS servers and back-end servers. In this way, we have relaxed hot spot of dispatcher. At the same time, we have preserved the balanced load state through estimation of client domain load rate. We have shown that the adaptive dispatching method improves overall performance in realistic workload simulation.

Mirinae: A peer-to-peer overlay network for large-scale content-based publish/subscribe systems

Content-based publish/subscribe systems provide a useful alternative to traditional address-based communication due to their ability to decouple communication between participants. It has remained a challenge to design a scalable overlay supporting the complexity of content-based networks, while satisfying the desirable properties large distributed systems should have. This paper presents the design of Mirinae, a new structured peer-to-peer overlay mesh based on the interests of peers. Given an event, Mirinae provides a flexible and efficient dissemination tree minimizing the participation of non-matching nodes. We also present a novel ID space transformation mechanism for balancing routing load of peers even with highly skewed data, which is typical of the real world. Mirinae can be used as a substrate for content-search and range query in other important distributed applications.

Least Popularity-Per-Byte Replacement algorithm for a proxy cache

With the recent explosion in usage of the World Wide Web, the problem of caching Web objects has gained considerable importance. The performance of these Web caches is highly affected by the replacement algorithm. Today, many replacement algorithms have been proposed for Web caching and these algorithms use the on-line fashion parameters. Recent studies suggest that the correlation between the on-line fashion parameters and the object popularity in the proxy cache are weakening due to the efficient client caches. We suggest a new algorithm, called Least Popularity Per Byte Replacement (LPPB-R). We use the popularity value as the long-term measurements of request frequency to make up for the weak point of the previous algorithms in the proxy cache and vary the popularity value by changing the impact factor easily to adjust the peformance to needs of the proxy cache. We examine the performance of this and other replacement algorithms via trace driven simulation.

An Energy-Efficient Scheduling MAC Protocol for Wireless Sensor Networks

This paper proposes a novel MAC protocol for sensor networks which is energy efficient and has a good performance as well. Energy is the most important resource in battery-operated sensor networks. Sensor nodes need the wireless network interface to forward data and it has to be always awake to handle possible traffics. This occupies most energy wastes in sensor nodes and pervious works solved this problem by periodic listen and sleep. However, periodic listen and sleep causes the loss of performance. Our protocol also uses periodic sleeping but listen period are used for multi-hop schedule reservations and data transmission is delayed until schedule time. Schedules are reserved as data are pipelined for energy efficiency and performance. In addition, we suggest an energy efficient error recovery mechanism for data transmissions. We evaluate our protocol and compare with previous MAC protocols through ns-2 simulations.

Reducing Replication Overhead for Data Durability in DHT Based P2P System

DHT based p2p systems appear to provide scalable storage services with idle resource from many unreliable clients. If a DHT is used in storage intensive applications where data loss must be minimized, quick replication is especially important to replace lost redundancy on other nodes in reaction to failures. To achieve this easily, a simple replication method directly uses a consistent set, such as a leaf set and a successor list. However, this set is tightly coupled to the current state of nodes and the traffic needed to support this replication can be high and bursty under churn. This paper explores efficient replication methods that only glimpse a consistent set to select a new replica. Replicas are loosely coupled to a consistent set and we can eliminate the compulsory replication under churn. Because of a complication of the new replication methods, the careful data management is needed under churn for the correct and efficient data lookup. Results from a simulation study suggest that our methods can reduce network traffic enormously for high data durability.

Efficient and Scalable Client-Clustering for Proxy Cache

Many cooperated web cache systems and protocols have been proposed. These systems, however, require expensive resources, such as external bandwidth and proxy cpu or storage, while inducing hefty administrative costs to achieve adequate client population growth. Moreover, a scalability problem in the cache server management still exists.