• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.17 no.2
• /
• pp.103-113
• /
• 2013

We analyze the queueing model for leaky bucket (LB) scheme with dynamic arrivals and token rates. In other words, in our LB scheme the arrivals and token rates are changed according to the buffer occupancy. In telecommunication networks, the LB scheme has been used as a policing function to prevent congestion. By considering bursty and correlated properties of input traffic, the arrivals are assumed to follow a Markov-modulated Poisson process (MMPP). We derive the distribution of system state, and obtain the loss probability and the mean waiting time. The analysis is done by using the embedded Markov chain and supplementary variable method. We also present some numerical examples to show the effect of our proposed model.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.65-65
• /
• 2000

For Video on Demand(VoD) servers, a design of an efficient scheduler is important to the support a large number of clients having various playback speeds and receiving rates. In this paper, we propose the scheduling algorithm to handle establishing deadlines and selection using the earliest deadline first. To establish deadlines and selections, the period of the receiving rates for each client is located between the over-max receiving rate and the over-playback rate. To avoid video starvation and the buffer overflow of each client, the proposed algorithm guarantees providing the admission control. Because of establishing deadlines and selection, period of each client receiving is between one over max receiving rate and one over play back rate. Using Virtual Buffer in server, scheduling load is reduced. The efficiency of the proposed algorithm is verified using a Petri Net_Based simulation tool, Exspect.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.38 no.8
• /
• pp.30-41
• /
• 2001

In this paper, the architecture of Dynamic GSMP V3(General Switch Management Protocol Version 3), an open interface protocol with resource management functions for efficient IP service on ATM over MPLS, is proposed and implemented in hardware. And we compare and analyze the proposed GSMP with the GSMP under standardization process in terms of CLR (Cell Loss Rate). We design the Slave block of the Dynamic GSMP V3 using SAM-SUNG SoG $0.5{\mu}m$ process, which performs functions for switch connection control in the ATM Switch. In order to compare difference performanaces between the proposed method and the conventional one, we conducts simulations using the minimum buffer search algorithm with random cell generation. The exponential results show that the proposed method leads to performance enhancement in CLR.

• Journal of the Korea Society for Simulation
• /
• v.14 no.1
• /
• pp.33-42
• /
• 2005

Service Level Agreement (SLA) is a service providing scheme by a service class agreement between a user and a service provider SLA allows that a user can select an expected service class in various service classes provided from a service provider. Recently, SLA management is adapted to support the end-to-end Qos for service users in Beyond 3 Generation (B3G) networks. In B3G networks, SLA provides multiple service classes on access networks so that service classes should be managed to assure the service satisfaction for users. In this paper, we propose a dynamic Qos management scheme by IP traffic class controlling based on SLA in B3G networks. Also, to manage dynamic traffic service, we consider Differentiated services (Diffserv) mechanism for the resource management by SLA. An If service traffic class on SLA can be dynamically changed by Diffserv traffic management to support dynamic end-to-end Qos. Hence, in this paper, we consider the buffer threshold scheme for controlling traffic loads and the traffic level control scheme for implementing the dynamic traffic management by the SLA.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.2
• /
• pp.1-8
• /
• 2013

TIn this paper, we propose a high performance L1 cache structure on the high clock CPU of 4GHz. The proposed cache memory consists of three parts, i.e., a direct-mapped cache to support fast access time, a two-way set associative buffer to exploit temporal locality, and a buffer-select table. The most recently accessed data is stored in the direct-mapped cache. If a data has a high probability of a repeated reference, when the data is replaced from the direct-mapped cache, the data is selectively stored into the two-way set associative buffer. For the high performance and low power consumption, we propose an one way among two ways set associative buffer is selectively accessed based on the buffer-select table(BST). According to simulation results, Energy $^*$ Delay product can improve about 45%, 70% and 75% compared with a direct mapped cache, a four-way set associative cache, and a victim cache with two times more space respectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.11B
• /
• pp.1016-1022
• /
• 2003

The EPON has been actively studied as one of the access networks for the economic configuration of FTTH. The EPON must support the dynamic bandwidth allocation to the subscribers in order to support the QoS due to its base on Ethernet technology EFM SG, which is actively working for the standardization of EPON, also recently decided to select DBA. Therefore in this paper, we designed a ONU buffer scheduling algorithm(AIWFQ) and a scheme of DBA(Class-based FCFS) for the OLT suitable for embodying MPCP of the EPON. In this paper, we proposed methods that the EPON system can make use of by measuring end to end process delay time and the buffer size in order to implement the algorithm by using the OPNET.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.2B
• /
• pp.272-278
• /
• 2004

The EPON has been actively studied as one of the access networks for the economic configuration of FTTH. The EPON must support the dynamic bandwidth allocation to the subscribers in order to support the QoS due to its base on Ethernet technology. EFM SG, which is actively working for the standardization of EPON, also recently decided to select DBA. Therefore in this paper, we designed a ONU buffer scheduling algorithm (AIWFQ) and a scheme of DBA(Class-based FCFS) for the OLT suitable for embodying MPCP of the EPON. In this paper, we proposed methods that the EPON system can make use of by measuring end to end process delay time and the buffer size in order to implement the algorithm by using the OPNET.

• Journal of IKEEE
• /
• v.5 no.1 s.8
• /
• pp.100-110
• /
• 2001

In this paper, we investigated several buffer management schemes for the design of shared-memory type ATM switches, which can enhance the utilization of switch resources and can support quality-of-service (QoS) functionalities. Our results show that dynamic threshold (DT) scheme demonstrate a moderate degree of robustness close to pushout(PO) scheme, which is known to be impractical in the perspective of hardware implementation, under various traffic conditions such as traffic loads, burstyness of incoming traffic, and load non-uniformity across output ports. Next, we considered buffer management strategies to support QoS functions, which utilize parameter values obtained via connection admission control (CAC) procedures to set tile threshold values. Through simulations, we showed that the buffer management schemes adopted behave well in the sense that they can protect regulated traffic from unregulated cell traffic in allocating buffer space. In particular, it was observed that dynamic partitioning is superior in terms of QoS support than virtual partitioning.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.38 no.11
• /
• pp.30-40
• /
• 2001

In this paper, we propose a dynamic slot allocation algorithm for efficient transmission of VBR services in wireless ATM networks. The proposed algorithm is based on a hybrid dynamic parameter(DP) control which combines the strength of in-band control and out-of-band control by considering the variation characteristics of buffer length in distributed mobile terminals. This algorithm consists of four sub-algorithms: dynamic parameter determination algorithm, dynamic parameter transmission algorithm, estimation algorithm of the number of request slots, and prorated-allocation algorithm. As the proposed allocation algorithm based on the hybrid DP control scheme can offer nearly precise MAC level estimations of the requirements for each VBR, the algorithm makes it possible to obtain ideal allocation efficiency. The allocation efficiency of the algorithm is shown by numerical analysis. Simulation results show that the proposed algorithm has better performance than conventional schemes in terms of allocation efficiency, delay and cell loss ratio under VBR traffic.

• ETRI Journal
• /
• v.35 no.1
• /
• pp.146-149
• /
• 2013

We propose a reconciliation sublayer (RS)-based lane and traffic control protocol for energy-efficient 40-G/100-G Ethernet. The RS performs active/inactive lane control and data rate adaptation depending on active lane information received from the upper layer. This protocol does not result in a processing delay in the media access control layer, nor is an additional buffer required at the physical layer for dynamic lane control. It ensures minimal delay and no overhead for the exchange of control frames and provides a simple adaptive data rate.