• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.4C
• /
• pp.295-307
• /
• 2002

The IEEE 1394 protocol is a de facto standard in multimedia digital home network. It supports several advanced features such as hot plugging, dynamic network reconfiguration, isochronous transmission and so on. Since the IEEE 1394 was adapted in the field of multimedia service with QoS guarantee, back bone network protocol with reel-time digital instrumentation and control sub networks, and physical layer protocol for real-time middleware such as real-time CORBA, the additional real-time features has been required in device driver implementation because of the necessity of the predictability enhancement. To guarantee the real-time features, the device driver of the IEEE 1394 should support the priority based packet processing and also need a isochronous buffer management mechanism to deal with the periodic isochronous communication. In this paper, we proposed a new software architecture of the IEEE 1394 device driver for supporting the real-time characteristics such as priority based packet processing, priority based scheduling and so on.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.4
• /
• pp.750-760
• /
• 1994

In this paper we proposed a multiplexing scheme of real time and non-real traffics in which a congestion control is embedded. Real time traffics are assumed to be nonqueuable and have preemptive priority over non-real time traffics in seizing the common output link, whereas the non-real time traffics wait in the common buffer if the output link is not available for transmission. Real time traffics are encoded according to the bandwidth reduction strategy, paticularly when congestion occurs among non-real time traffics. This scheme provides us an efficient way for utilizing the costly bandwidth resources, by accommodation as many real time traffics as possible with gauranteeing its mimimum bandwidth requirements, and also resloving the congestion encountered among non-real time traffics. We describe the system as a Markov queueing system, provide the analysis by exploiting the matrix geometric method, and present the performance for various performance measures of interest. Some numerical results are also provided.

• Journal of the Korea Society for Simulation
• /
• v.29 no.1
• /
• pp.57-69
• /
• 2020

Many researchers have recommended that DBR scheduling would be an efficient method to maintain the balance of their workload among many processes in the general flow shop. However, as product variety has increased in recent years, the process has become more complex and requires the re-entrance of raw materials and work in process. The re-entrant line has known for the complex manufacturing process that raw materials are repeatedly processed on the same machine. This study reviews the applicability of DBR against the re-entrant manufacturing line due to the distinguishing characteristics and the higher complexity caused by multiple visits of a job into the identical process. In order to apply the DBR method to the re-entrant process, the main idea is to reconstruct re-entrant process into a virtual flow process(loop) that has a single bottleneck. This study discusses the following two questions. First, DBR is also superior to traditional scheduling methods against re-entrant manufacturing line. And how we structure and detect the system bottleneck (or sub-bottleneck) through drum-buffer-rope concepts. To answer the above questions, we experimented and analyzed the effects of the applicability of DBR under the general re-entrant process model(TRC, Technology Research Center). As a result, we have identified a balance between loops for cycle time and work in process.

• The KIPS Transactions:PartA
• /
• v.10A no.6
• /
• pp.691-700
• /
• 2003

Due to the fast development of wire&wireless internet and computer hardware, more and more internet services are being developed, such as Internet broadcast, VoD (Video On Demand), etc. So QoS (Qualify of Service) is essentially needed to guarantee the quality of these services. Traditional Internet is Best-Effort service in which all packets are transported in FIFO (First In First Out) style. However, FIFO is not suitable to guarantee the quality of some services, so more research in QoS router and QoS protocol are needed. Researched QoS router and protocol are high cost and inefficient because the existing infra is not used. To solve this problem, a new QoS control method, named Network Adaptive QoS, is introduced and applied to client/server-based streaming systems. Based on network bandwidth monitoring mechanism, network adaptive QoS control method can be used in wire&wireless networks to support QoS in real-time streaming system. In order to reduce application cost, the existing streaming service is used in NAQoS. A new module is integrated into the existing server and client. So the router and network line are not changed. By simulation in heavy traffic network conditions, we proved that stream cannot be seamless without network adaptive QoS method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.18 no.5
• /
• pp.708-720
• /
• 1993

The new preventive control mechanisms for traffic management in BISDN/ATM networks can be divided into Connection Admission Control(CAC), Usage Parameter Control (UPC), and Priority Control. Of these mechanism, Usage Parameter Control continuously monitors the parameters admitted in the network's entry point to guarantee quality of service of connections already admitted. Upon detecting traffic that violates the negotiated parameter, it takes the necessary control measures to prevent congestion. Among these traffic control methods, this paper focuses on the Usage Parameter Control method, and proposes and designs GRACE-LB(Guaranteed Rate Acceptance & Control Element-using Leaky Bucket) which improves upon existing UPC models. GRACE-LB modifies the previous LB model by eliminating the cell buffer, dividing the token Pool into two pools, Long-term pool, Short-term pool, and changing the long-term token generating form using 'Cycle Token' into the same bursty form as the traffic source. Through this, GRACE-LB achieves effective control of the Average Bit Rate(ABR) and burst duration of bursty multimedia traffic which previous LB models found difficult to control. Also, since GRACE-LB can e implemented using only simple operations and there are no cell buffers in it, it has the merit of being easily installed at any place.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7B
• /
• pp.613-627
• /
• 2004

In this paper, we propose an Aggregate Fairness Maker (ARM) required for an Edge router to improve fairness of throughput among the flows of Assured Service in DiffServ with different round trip time (RTT) and we propose a user flow Three Color Marker (uf-TCM) as a flow marker that marks packets from the flow as green, yellow, or red. A yellow packet is the packet that consumes loss token in uf-TCM as well as that is demoted green packet in AM due to disobey the aggregate traffic profile. The proposed AFH promotes yellow packet to green packet or demotes green packet to yellow packet through the fair method without per-flow management, and it improves the feirness of throughput among the flows as well as link utilization. A yellow packet and a red packet have the same drop precedence at Core Router in our scheme. So we can use the RIO buffer management scheme. We evaluated the performance of our proposed AFM and the REDP Marker that was proposed to improve fairness without per-flow management. Simulation results show that, compared with the REDP marker, proposed AFM can improve performance of throughput fairness among the flows with different RTT and link utilization under the over-provisioning, exact-provisioning, and under-provisioning network environments at Multiple DiffServ domains as well as at Single DiffServ domain.