Visual Place Recognition (VPR) aims to retrieve frames from a geotagged database that are located at the same place as the query frame. To improve the robustness of VPR in perceptually aliasing scenarios, sequence-bas...
详细信息
Data dependency, often presented as directed acyclic graph (DAG), is a crucial application semantics for the performance of data analytic platforms such as Spark. Spark comes with two built-in schedulers, namely FIFO ...
详细信息
ISBN:
(数字)9781728168760
ISBN:
(纸本)9781728168777
Data dependency, often presented as directed acyclic graph (DAG), is a crucial application semantics for the performance of data analytic platforms such as Spark. Spark comes with two built-in schedulers, namely FIFO and Fair scheduler, which do not take advantage of data dependency structures. Recently proposed DAG-aware task scheduling approaches, notably GRAPHENE, have achieved significant performance improvements but paid little attention to cache management. The resulted data access patterns interact poorly with the built-in LRU caching, leading to significant cache misses and performance degradation. On the other hand, DAG-aware caching schemes, such as Most Reference Distance (MRD), are designed for FIFO scheduler instead of DAG-aware task *** this paper, we propose and develop a middleware Dagon, which leverages the complexity and heterogeneity of DAGs to jointly execute task scheduling and cache management. Dagon relies on three key mechanisms: DAG-aware task assignment that considers dependency structure and heterogeneous resource demands to reduce potential resource fragmentation, sensitivity-aware delay scheduling that prevents executors from long waiting for tasks insensitive to locality, and priority-aware caching that makes the cache eviction and prefetching decisions based on the stage priority determined by DAG-aware task assignment. We have implemented Dagon in Apache Spark. Evaluation on a testbed shows that Dagon improves the job completion time by up to 42% and CPU utilization by up to 46% respectively, compared to GRAPHENE plus MRD.
With the emerging of smart metering around the world, there is a growing demand to analyse the residential energy usage. In this paper, we propose a Deep Neural Network (DNN)-based approach for non-intrusive load moni...
详细信息
Thanks to the large-scale smart meters deployments around the world, non-intrusive appliance load monitoring (NILM) is receiving popularity. It aims to disaggregate the total electricity load of a home into individual...
详细信息
Millimeter wave communication provides high data rates thanks to large arrays at the transmitter and receiver, coupled with large bandwidth channels. Exploiting the arrays is challenging due to the need to configure p...
详细信息
We address a special kind of Internet of Things (IoT) systems that are also real-time. We call them real-time IoT (RT-IoT) systems. An RT-IoT system needs to meet timing constraints of system delay, clock synchronizat...
详细信息
We address a special kind of Internet of Things (IoT) systems that are also real-time. We call them real-time IoT (RT-IoT) systems. An RT-IoT system needs to meet timing constraints of system delay, clock synchronization, deadline, and so on. The timing constraints turn to be more stringent as we get closer to the physical things. Based on the reference architecture of IoT (ISO/IEC 30141), the RT-IoT conceptual model is established. The idea of edge subsystem is introduced. The sensing & con-trolling domain is the basis of the edge subsystem, and the edge subsystem usually must meet the hard real-time constraints. The model includes four perspectives, the time view, computation view, communication view, and control view. Each view looks, from a different angle, at how the time parameters impact an RT-IoT system.
Modern datacenter schedulers apply a static policy to partition resources among different tasks. The amount of allocated resource won't get changed during a task's lifetime. However, we found that resource usa...
详细信息
ISBN:
(纸本)9781538641293
Modern datacenter schedulers apply a static policy to partition resources among different tasks. The amount of allocated resource won't get changed during a task's lifetime. However, we found that resource usage during a task's runtime demonstrates high dynamics and it only reaches full usage at few moments. Therefore, the static allocation policy doesn't exploit the dynamic nature of resource usage, leading to low system resource utilization. To address this hard problem, a recently proposed task-consolidation approach packs as many tasks as possible on the same node based on real-time resource demands. However, this approach may cause resource over-allocation and harm application performance. In this paper, we propose and develop ECS, an elastic container based scheduler that leverages resource usage variation within the task lifetime to exploit the potential utilization and parallelism. The key idea is to proactively select and shift tasks backward so that the inherent paralleled tasks can be identified without over-allocation. We formulate the scheduling scheme as an online optimization problem and solves it using a resource leveling algorithm. We have implemented ECS in Apache Yarn and performed evaluations with various MapReduce benchmarks in a cluster. Experimental results show that ECS can efficiently utilize resource and achieves up to 29% reduction on average job completion time while increasing CPU utilization by 25%, compared to stock Yarn.
The deep two-stream architecture [23] exhibited excellent performance on video based action recognition. The most computationally expensive step in this approach comes from the calculation of optical flow which preven...
详细信息
ISBN:
(纸本)9781467388511
The deep two-stream architecture [23] exhibited excellent performance on video based action recognition. The most computationally expensive step in this approach comes from the calculation of optical flow which prevents it to be real-time. This paper accelerates this architecture by replacing optical flow with motion vector which can be obtained directly from compressed videos without extra calculation. However, motion vector lacks fine structures, and contains noisy and inaccurate motion patterns, leading to the evident degradation of recognition performance. Our key insight for relieving this problem is that optical flow and motion vector are inherent correlated. Transferring the knowledge learned with optical flow CNN to motion vector CNN can significantly boost the performance of the latter. Specifically, we introduce three strategies for this, initialization transfer, supervision transfer and their combination. Experimental results show that our method achieves comparable recognition performance to the state-of-the-art, while our method can process 390.7 frames per second, which is 27 times faster than the original two-stream method.
As a powerful analysis tool of Petri nets, reachability trees are fundamental for systematically investigating many characteristics such as boundedness, liveness and reversibility. This work proposes a method to gener...
详细信息
In many distributed environments, multiple processes both interact/collaborate with each other and share some common resources. To model and analyze such systems, this paper defines a class of Petri nets called Parall...
详细信息
ISBN:
(纸本)9781467399760
In many distributed environments, multiple processes both interact/collaborate with each other and share some common resources. To model and analyze such systems, this paper defines a class of Petri nets called Parallel Processes Net (P 2 N). A P 2 N composes of a group of Single-process Nets (SNs). Each SN models a process and these SNs are connected via a set of places. Some of these places represent the common resources shared by these processes, and others represent the channels through which messages are transferred among these processes. We define collaborative-ness for P 2 Ns which requires that each process of a modeled system never enters into a deadlock or livelock state. The collaborative-ness can be represented by a CTL (Computational Tree Logic) formula and thus can be checked (for bounded P 2 Ns) by using some Petri net tools such as INA.
暂无评论