scispace - formally typeset
Search or ask a question
Topic

Scalability

About: Scalability is a research topic. Over the lifetime, 50930 publications have been published within this topic receiving 931614 citations.


Papers
More filters
Journal ArticleDOI
TL;DR: System-C based simulation result shows that, compared with diagonal Mesh (DMesh), diagonal Torus (DTorus) and XMesh networks, the SD-Torus network can achieve high performance with a lower network cost, making it a powerful candidate for the high performance interconnection networks.

21 citations

Journal ArticleDOI
TL;DR: This paper proposes an asynchronous mechanism to inform the kernel that a buffer is ready to read, which ensures that tracing sites do not require any kernel primitive, and therefore protects from infinite recursion.
Abstract: Studying execution of concurrent real-time online systems, to identify far-reaching and hard to reproduce latency and performance problems, requires a mechanism able to cope with voluminous information extracted from execution traces. Furthermore, the workload must not be disturbed by tracing, thereby causing the problematic behavior to become unreproducible.In order to satisfy this low-disturbance constraint, we created the LTTng kernel tracer. It is designed to enable safe and race-free attachment of probes virtually anywhere in the operating system, including sites executed in non-maskable interrupt context.In addition to being reentrant with respect to all kernel execution contexts, LTTng offers good performance and scalability, mainly due to its use of per-CPU data structures, local atomic operations as main buffer synchronization primitive, and RCU (Read-Copy Update) mechanism to control tracing.Given that kernel infrastructure used by the tracer could lead to infinite recursion if traced, and typically requires non-atomic synchronization, this paper proposes an asynchronous mechanism to inform the kernel that a buffer is ready to read. This ensures that tracing sites do not require any kernel primitive, and therefore protects from infinite recursion.This paper presents the core of LTTng's buffering algorithms and measures its performance.

21 citations

Journal ArticleDOI
TL;DR: This paper investigates several IoT search scenarios and proposes a uniform representation model for sensor information recordings and develops information retrieval architecture for IoT, where an indexing mechanism called efficiency maximization and cost minimization is proposed to solve the property selection problem in the process of index construction and update.
Abstract: Billions of devices are connected in the Internet of Things (IoT)-based sensor networks and they continuously generate a large volume of data. In order to get access to specific data, which is crucial to enable a myriad of new intelligent applications, efficient information retrieval becomes an imminent need for IoT. However, sensor information in the physical world can be heterogeneous, high dimensional, and voluminous due to the complex and dynamic environments. In this paper, we first investigate several IoT search scenarios and propose a uniform representation model for sensor information recordings. Four query models are designed to represent all possible information query styles. With these models, we develop information retrieval architecture for IoT. In essence, an indexing mechanism called efficiency maximization and cost minimization is proposed to solve the property selection problem in the process of index construction and update. Meanwhile, a novel real-time grid R-tree structure is designed to support historical and real-time search for spatiotemporal observation data. Simulation results based on real-world IoT data sets show that storage space is considerably reduced with the sensor model. Furthermore, the proposed indexing mechanisms can improve retrieval efficiency and accuracy, and ensure scalability for large-sized data simultaneously.

21 citations

Proceedings ArticleDOI
01 Jan 2019
TL;DR: This algorithm, the Fast Density-Grid Clustering Algorithm, works by dividing the data space into a grid structure and then assigning a density measurement to each grid cell in order to cluster the entire space.
Abstract: Clustering algorithms are a large area of focus in the world of data analytics, especially in the current age of big data sets that are large, multidimensional, and grow quickly. We also live in an age where parallel and distributed computing are at the forefront of any form of data analysis. Because of these two ideas, this paper discusses an algorithm that was designed to scale well with big data sets for fast run time, while also being highly scalable for a potential parallel version. This algorithm, the Fast Density-Grid Clustering Algorithm, works by dividing the data space into a grid structure and then assigning a density measurement to each grid cell. These spaces are then merged based on their densest neighbor in order to cluster the entire space. The experimental results for the serial version demonstrate that the algorithm is generally comparable to DBSCAN in accuracy, while also having a lower run time.

21 citations

Journal ArticleDOI
TL;DR: The growing volume of data produced continuously in the Cloud and at the Edge poses significant challenges for large-scale AI applications to extract and learn useful information from the data in a...
Abstract: The growing volume of data produced continuously in the Cloud and at the Edge poses significant challenges for large-scale AI applications to extract and learn useful information from the data in a timely and efficient way. The goal of this article is to explore the use of computational storage to address such challenges by distributed near-data processing. We describe Newport, a high-performance and energy-efficient computational storage developed for realizing the full potential of in-storage processing. To the best of our knowledge, Newport is the first commodity SSD that can be configured to run a server-like operating system, greatly minimizing the effort for creating and maintaining applications running inside the storage. We analyze the benefits of using Newport by running complex AI applications such as image similarity search and object tracking on a large visual dataset. The results demonstrate that data-intensive AI workloads can be efficiently parallelized and offloaded, even to a small set of Newport drives with significant performance gains and energy savings. In addition, we introduce a comprehensive taxonomy of existing computational storage solutions together with a realistic cost analysis for high-volume production, giving a good big picture of the economic feasibility of the computational storage technology.

21 citations


Network Information
Related Topics (5)
Network packet
159.7K papers, 2.2M citations
91% related
Wireless sensor network
142K papers, 2.4M citations
90% related
Wireless network
122.5K papers, 2.1M citations
88% related
Wireless
133.4K papers, 1.9M citations
88% related
Cluster analysis
146.5K papers, 2.9M citations
88% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
20236,133
202212,865
20212,854
20203,111
20193,218
20183,140