• 인터넷정보학회논문지
• /
• 제20권5호
• /
• pp.113-119
• /
• 2019

최근 상용화된 5G 이동통신의 주요한 특징은 High Data Rate와 Connection Density 그리고 Low Latency로 대표할 수 있는데, 이중 기존 4G와 가장 차별되는 특징은 Low Latency로 다양한 새로운 서비스 제공의 기반이 될 것이다. 이러한 특징을 활용한 서비스로는 AR, 자율주행 등이 검토되고 있으며 관련표준에서도 5G Network Latency 논의를 진행하고 있다. 그러나 서비스 관점의 E2E Latency 논의는 많이 부족한 것이 사실이다. 5G에서 Low Latency를 달성을 위한 최종목표는 RTD 기준 Air Interface 1ms 달성으로 이는 '20년 초 Rel-16을 통한 URLLC(Ultra-reliable Low Latency Communications)를 통해 가능하며, 추가적으로 MEC(Moble Edge Computing)를 통한 Network latency 감소도 연구 중이다. 전체 5G E2E Latency는 5G Network 관련 외에도 다양한 요인이 존재하는데, 주요 요인으로는 5G Network과 서비스 제공을 위한 IDC Server 사이의 경로에 놓인 선로/장비 Latency, 단말 App과 Server 내 서비스 처리를 위한 Processing Latency 등이 존재한다. 한편, 서비스 초기 Setup을 위한 Latency와 서비스가 지속 중인 경우의 Latency를 구분하여 세부 서비스 요구사항에 대하여 연구하는것도 필요한데, 이를 위해 본 논문에서는 서비스 초기 Setup과 관련하여 다음과 같은 세가지 요인에 대하여 검토를 진행하였다. 첫째로 (1) Data호 Setup시에 발생 가능한 Latency, 둘째 전력 효율화를 위한 (2) CRDX On/Off에 따른 영향, 마지막으로 (3) H/O가 발생되는 경우에 Latency에 대하여 Latency에 미치는 영향을 실험과 분석을 제시했다. 이를 통해 우리는 Low Latency가 필요한 서비스의 초기 Setup시에 Latency와 관련된 서비스 요구사항 및 기획에 기여할 수 있을 것으로 기대한다.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2012년도 제46차 하계학술발표논문집 20권2호
• /
• pp.297-300
• /
• 2012

현재 cache의 성능 향상을 위한 많은 기법들이 제안되고 있으며, Latency Hiding 기법 역시 cache의 효율적인 사용을 위해 많은 연구가 진행 되어 왔다. write buffer를 사용한 write Latency hiding기법이나 multi threading을 사용한 Latency Hiding 방법 등 여러 기법들이 연구되어 왔으며, 지금도 Latency hiding을 위한 많은 연구들이 지속적으로 진행되고 있다. 본 논문 역시 효율적인 Latency Hiding을 위한 누적 버퍼를 제안한다. 본 논문은 누적 버퍼의 활용도를 조사하여 얼마나 효율적으로 Latency를 은폐했는지, 또 버퍼를 사용함으로써 얻는 다른 이점에 대해 집중적으로 연구하였다.

• 한국통신학회논문지
• /
• 제30권12B
• /
• pp.847-854
• /
• 2005

패킷 스위칭 네트워크 노드에서 정해진 대역폭을 보장해주는 스케줄링 기법 중 많은 수가 Latency-Rate (LR) server의 범주에 포함된다. LR server로 이루어진 네트워크는 각 노드에서의 latency의 합이 그대로 전체 네트워크의 latency가 된다는 성질을 가지고 있어서 각 노드의 latency를 구하면 전체 네트워크에서 겪을 수 있는 최대 delay를 쉽게 얻을 수 있다. Deficit Round Robin (DRR)은 이러한 LR server의 하나로 가장 구현이 간단하다는 장점 때문에 실제 여러 종류의 시스템에서 채택하였다. 본 연구에서는 이러한 DRR을 개선한 "DRR with Instant Service" (DRR-IS) 기법을 제시하고 분석하였다. DRR-IS가 LR server임을 증명하였고, 기존 DRR과 비교하며 전혀 복잡도의 증가가 없으면서 latency 성능은 $30\%$ 가량 향상되었음을 밝혀내었다.

• 한국인터넷방송통신학회논문지
• /
• 제17권1호
• /
• pp.33-39
• /
• 2017

차량 네트워크, 분산 로봇과 사이버 물리 시스템 등 mission-critical 환경에서 동작하는 서비스들은 기존 서비스에 비해 latency에 대한 요구사항이 훨씬 엄격하다. 그 중 자율주행 차량 간 통신은 급격히 부상하고 있는 애플리케이션 영역으로 수 ms 수준의 엄격한 latency 요구사항을 가진다. 802.11p나 LTE-direct standards를 이용하는 현재 시스템은 이러한 수준의 ultra-low latency를 만족하지 못한다. 현재, 전체 latency의 상당부분을 medium access가 차지하기 때문에, 이 부분의 해결을 위해서는 Layer2를 수정할 필요가 있다. 따라서 본 논문에서는 MAC layer에 초점을 맞추어 coordinator에 의한 scheduling 없이 join/leave가 자유로운 Autonomous TDMA(ATDMA)를 고안하여 scalable하면서 latency가 보장되는 MAC을 구현하였다. 또한 WAVE 프로토콜과의 비교를 통해 알고리즘의 성능을 평가하였다.

• Annals of Clinical Neurophysiology
• /
• 제4권2호
• /
• pp.108-113
• /
• 2002

Background and Objective : Carpal tunnel syndrome (CTS) is the most common mononeuropathy encountered in clinical practice. No single procedure or group of procedures has demonstrated adequate sensitivity. F-wave study in CTS is very rarely reported. To determine the diagnostic usefulness of new parameters of F-wave and comparative study of F-wave parameters of median and ulnar nerves in patients with CTS. Methods : F-wave responses of median and ulnar nerves were analyzed from 27 patients with clinically diagnosed CTS and 22 age and gender-matched normal control. Conventional F-wave parameters were studied. Also, the usefulness of new parameters such as mean and maximal ulnar-median F-wave latency differences, ulnar-median F-wave persistence and chronodispersion differences, median/ulnar F-wave amplitude ratio, and F-wave conduction velocity (FCV) using mean and maximal latency were assessed. Results : Compared with controls, median F-wave minimal, maximal and mean latencies, mean F-wave amplitude/M-wave amplitude, minimal, mean and maximal ulnar-median F-wave latency differences, and FCVs using minimal, maximal and mean latency were significant (P<0.05~0.001). Median F-wave minimal, maximal and mean latencies, mean ulnar-median F-wave latency difference, and FCVs using minimal, maximal and mean latency showed high sensitivity and specificity. Mean ulnar-median F-wave latency difference and FCVs using maximal and mean latency were new parameters. Conclusion : New F-wave parameter including mean ulnar- median F-wave latency difference and FCVs using maximal and mean latency may be a useful to assess the CTS. Also, median F-wave minimal, maximal and mean latencies, and FCV using minimal latency may be included in routine diagnostic tests in CTS.

• Journal of Positioning, Navigation, and Timing
• /
• 제6권1호
• /
• pp.17-26
• /
• 2017

Latency occurs in RTK, where the measured position actually outputs past position when compared to the measured time. This latency has an adverse effect on the navigation accuracy. In the present study, a system that estimates the latency of RTK and compensates the position error induced by the latency was implemented. To estimate the latency, the speed obtained from an odometer and the speed calculated from the position change of RTK were used. The latency was estimated with a modified correlator where the speed from odometer is shifted by a sample until to find best fit with speed from RTK. To compensate the position error induced by the latency, the current position was calculated from the speed and heading of RTK. To evaluate the performance of the implemented method, the data obtained from an actual vehicle was applied to the implemented system. The results of the experiment showed that the latency could be estimated with an error of less than 12 ms. The minimum data acquisition time for the stable estimation of the latency was up to 55 seconds. In addition, when the position was compensated based on the estimated latency, the position error decreased by at least 53.6% compared with that before the compensation.

• Journal of Multimedia Information System
• /
• 제5권1호
• /
• pp.59-62
• /
• 2018

The delay between input action and visual interface feedback ("Latency") in a touchscreen inking task reduces the user's performance. When the latency is less than 2.38ms, the user cannot perceive the latency in dragging task. This value is difficult to achieve on recent touchscreens and general purpose computers. So, methods of predicting touch points to reduce perceptible latency has been proposed. In general, touch points prediction is not perfect. When using point prediction, feedback of the predicted points is displayed on the screen, after a while, erased when the actual points are displayed. When this task is implemented by software, it causes additional latency to work to erase predicted points feedback. It therefore propose a platform for rendering point prediction feedback without additional latency by the FPGA. This platform transmits input points and HDMI signals rendering feedback of input points to the FPGA. The FPGA draws the feedback of points predicted based on the input points on the HDMI and displays the screen. Since hardware rendering changes the HDMI signal every frame, it does not require erasing work and rendering can be done within an early time regardless of the amount of rendering, so we will reduce the latency.

• 음성과학
• /
• 제7권1호
• /
• pp.77-87
• /
• 2000

Distortion Product Otoacoustic Emissions (DPOAEs) can be measured in the external ear canal two fold: amplitude and latency, but most DPOAE studies deal with amplitude aspects. The purpose of this study was to investigate the latency of the 2f1-f2 DPOAEs in ears with hearing losses and to see if it could be a clinically useful method to distinguish normal from abnormal ears. For this purpose, DPOAE latency were measured as a function of frequency from 1 to 8 kHz in 30 ears with conductive and sensorineural hearing losses (SNHLs). DPOAEs were recorded with Otodynamic Analyzer ILO92. Results showed that the latency decreased as the frequency increased up to 8 kHz. The mean values of DPOAE latency for ears of SNHLs were shorter at all frequencies when they were compared to the mean values of normal ears. The latency in ears of conductive hearing losses was shorter than normal ears at the selective frequencies, as well. The results support the hypothesis that latency values are shorter in pathological ears.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권11호
• /
• pp.4016-4027
• /
• 2014

Understanding network latency is important for providing consistent and acceptable levels of services in network-based applications. However, due to the difficulty of estimating applications' network demands and the difficulty of network latency modeling the management of network resources has often been ignored. We expect that, since network latency repeats cycles of congested states, a systematic classification method for network status would be helpful to simplify issues in network resource managements. This paper presents a simple empirical method to classify network status with a real operational network. By observing oscillating behavior of end-to-end latency we determine networks' status in run time. Five typical network statuses are defined based on a long-term stability and a short-term burstiness. By investigating prediction accuracies of several simple numerical models we show the effectiveness of the network status classification. Experimental results show that around 80% reduction in prediction errors depending on network status.

• 한국컴퓨터정보학회논문지
• /
• 제20권9호
• /
• pp.1-10
• /
• 2015

In this paper, we propose a new variable latency L1 data cache architecture for multi-core processors. Our proposed architecture extends the traditional variable latency cache to be geared toward the multi-core processors. We added a specialized data structure for recording the latency of the L1 data cache. Depending on the added latency to the L1 data cache, the value stored to the data structure is determined. It also tracks the remaining cycles of the L1 data cache which notifies data arrival to the reservation station in the core. As in the variable latency cache of the single-core architecture, our proposed architecture flexibly extends the cache access cycles considering process variation. The proposed cache architecture can reduce yield losses incurred by L1 cache access time failures to nearly 0%. Moreover, we quantitatively evaluate performance, power, energy consumption, power-delay product, and energy-delay product when increasing the number of cache access cycles.