• 대한임베디드공학회논문지
• /
• 제16권1호
• /
• pp.9-15
• /
• 2021

Recently, it is increasingly necessary to run high-end neural network applications with huge computation overhead on top of resource-constrained embedded systems, such as wearable devices. While the huge computational overhead can be alleviated by distributed neural networks running on multiple separate devices, existing distributed neural network techniques suffer from a large traffic between the devices; thus are very vulnerable to communication failures. These drawbacks make the distributed neural network techniques inapplicable to wearable devices, which are connected with each other through unstable and low data rate communication medium like human body communication. Therefore, in this paper, we propose a distributed neural network partitioning technique that is resilient to communication failures. Furthermore, we show that the proposed technique also improves the inference accuracy even in case of no communication failure, thanks to the improved network partitioning. We verify through comparative experiments with a real-life neural network application that the proposed technique outperforms the existing state-of-the-art distributed neural network technique in terms of accuracy and resiliency to communication failures.

• 한국정보처리학회논문지
• /
• 제7권5호
• /
• pp.1426-1433
• /
• 2000

As distributed computing systems become popular, many modeling techniques and methods have been developed in specify the specification formally and verify the distributed/concurrent sytems. In spite of importance of communication in distributed object-oriented systems, specifying of communication method generally has not been emphasized in the design phase. One reason is due to the system designer misunderstanding, that is, a specification needs to be independent on the implementation. However, since defining communication pattern indistributed object-oriented systems is more serious than the required message passing method in the design phase. specifying the communication pattern is necessary instead of postponing until the implementation. In this paper, two formal specification techniques, temporal logic method and Petri nets method, for the communication in distributed systems are discussed. One is based on the temporal logic, which specifies the different patterns of primitive predicates. This method enable to define the underlying mechanism which can be interpreted as constraints. The Petri net method helps to specify the dynamic behavior of communicational patterns using the properties of Petri nets.

• 기술사
• /
• 제30권2호
• /
• pp.79-86
• /
• 1997

As network has high speed and wide communication capability, users demand diverse and new software to satisfy their needs. To meet users needs, the softwares for multimedia or groupware or distributed virtual environments can communicate the widely distributed information fast and accurately. Even though the technology for this is under development, it is deficient to support the reliable computer communication. Distributed Object System aims, as the new paradigm of distributed system software development to overcome this problem, to obtain in distributed environment the easiness of development and management, expandability, reusability which object oriented technologies support by solving the complexity of communication processing through the object oriented methods. This paper aims to introduce distributed object system, its technological properties and the current status and trend of technology development related to its standardization. Additionally, with explaining the Replicated Shared Object System(RSOS) which is developed in our country as one of the distributed object systems, its future prospects and technical issues are discussed.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권2호
• /
• pp.161-168
• /
• 2024

As the size of big data models grows, distributed training is emerging as an essential element for large-scale machine learning tasks. In this paper, we propose ParamHub for distributed data training. During the training process, this agent utilizes the provided data to adjust various conditions of the model's parameters, such as the model structure, learning algorithm, hyperparameters, and bias, aiming to minimize the error between the model's predictions and the actual values. Furthermore, it operates autonomously, collecting and updating data in a distributed environment, thereby reducing the burden of load balancing that occurs in a centralized system. And Through communication between agents, resource management and learning processes can be coordinated, enabling efficient management of distributed data and resources. This approach enhances the scalability and stability of distributed machine learning systems while providing flexibility to be applied in various learning environments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권6호
• /
• pp.2606-2626
• /
• 2016

This article investigates the distributed data storage problem in the space information network (SIN) using distributed fountain codes. Since space nodes in the SIN are resource-limited, in order to reduce energy consumption while improving the storage reliability, an efficient distributed storage based on fountain codes and probabilistic broadcasting (DSFPB) strategy is proposed. In the proposed strategy, source packets are disseminated among the entire network according to probabilistic broadcasting (PBcast), and the final degree distribution is close to the desired robust soliton distribution (RSD), this is benefited from the appropriate packets encoding procedure of the proposed strategy. As presented by the analysis and simulations, the total cost of data dissemination is greatly reduced compared with existing representative strategies, while improving the decoding performance.

• 한국정보통신학회논문지
• /
• 제16권3호
• /
• pp.460-469
• /
• 2012

본 논문에서는 UWB 기술 기반 WiMedia Distributed Medium Access Control (D-MAC) 표준 프로토콜에 적용할 수 있는 협력 통신 프로토콜을 제안한다. 이를 위해 무선 채널 상태에 따라 변화하는 UWB 링크 전송 속도에 적응적인 릴레이 노드 선정 알고리즘을 제안한다. 본 논문에서 제안하는 UWB 링크에 적응적인 릴레이 통신 프로토콜은 분산적인 D-MAC 표준 기술 및 무선 USB 기술과 호환성을 갖고, 각 디바이스에서 독립적으로 실행되는 Relay Node Selection (RNS) 기준에 따라 실행된다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권10호
• /
• pp.3771-3788
• /
• 2015

This paper studies the distributed energy efficient resource management in the Internet of Things (IoT). Wireless communication networks support the IoT without limitation of distance and location, which significantly impels its development. We study the communication channel and energy management in the wireless communication network supported IoT to improve the ability of connection, communication, share and collaboration, by using the game theory and distributed learning algorithm. First, we formulate an energy efficient neighbor collaborative game model and prove that the proposed game is an exact potential game. Second, we design a distributed energy efficient channel selection learning algorithm to obtain the global optimum in a distributed manner. We prove that the proposed algorithm will asymptotically converge to the global optimum with geometric speed. Finally, we make the simulations to verify the theoretic analysis and the performance of proposed algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권3호
• /
• pp.911-931
• /
• 2021

Recently, most cloud services use Docker container environment to provide their services. However, there are no researches to evaluate the performance of communication libraries for multi-GPU based distributed deep learning in a Docker container environment. In this paper, we propose an efficient communication architecture for multi-GPU based deep learning in a Docker container environment by evaluating the performances of various communication libraries. We compare the performances of the parameter server architecture and the All-reduce architecture, which are typical distributed deep learning architectures. Further, we analyze the performances of two separate multi-GPU resource allocation policies - allocating a single GPU to each Docker container and allocating multiple GPUs to each Docker container. We also experiment with the scalability of collective communication by increasing the number of GPUs from one to four. Through experiments, we compare OpenMPI and MPICH, which are representative open source MPI libraries, and NCCL, which is NVIDIA's collective communication library for the multi-GPU setting. In the parameter server architecture, we show that using CUDA-aware OpenMPI with multi-GPU per Docker container environment reduces communication latency by up to 75%. Also, we show that using NCCL in All-reduce architecture reduces communication latency by up to 93% compared to other libraries.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2011년도 한국컴퓨터종합학술대회논문집 Vol.38 No.1(D)
• /
• pp.213-214
• /
• 2011

This paper presents a distributed relay selection algorithm for cooperative communication. The algorithm separates the decision making into two simple steps, decision making for employing cooperative communication and decision making for relay selection.

• Transactions on Control, Automation and Systems Engineering
• /
• 제3권3호
• /
• pp.190-195
• /
• 2001

The advance of computer, network, and Internet technology enables the control systems to process the massive data in the distributed computing environments. To implement and maintain the software in distributed environment, the component-based methodology is widely used. This paper proposes the middleware architecture for the distributed computer control system. With the proposed middleware services, it is relatively easy to maintain compatibility between products and to implement a portable control application. To achieve the compatibility between heterogeneous systems, the proposed architecture provides the communication protocols based on the XML with lightweight event-based service.