Sining Wu

Delayed switching applied to memristor neural networks

Magnetic flux and electric charge are linked in a memristor. We reported recently that a memristor has a peculiar effect in which the switching takes place with a time delay because a memristor possesses a certain inertia. This effect was named the “delayed switching effect.” In this work, we elaborate on the importance of delayed switching in a brain-like computer using memristor neural networks. The effect is used to control the switching of a memristor synapse between two neurons that fire together (the Hebbian rule). A theoretical formula is found, and the design is verified by a simulation. We have also built an experimental setup consisting of electronic memristive synapses and electronic neurons.

Delayed Switching in Memristors and Memristive Systems

It was found that the switching in a memristor takes place with a time delay (this peculiar feature is named “the delayed switching”). This feature has been verified by a circuit-based experiment. The physical interpretation of this phenomenon is that an electron element possesses certain inertia, i.e., charge q or flux is inertial with the tendency to remain unchanged (settle to some equilibrium state). It cannot respond as rapidly as the fast variation in the excitation waveform and always takes a finite but small time interval to change its resistance value, as it must take place in a memristor or memristive system. In addition, a potential application of using this feature in ultradense computer memory has been discussed.

Grid-Oriented Storage: A Single-Image, Cross-Domain, High-Bandwidth Architecture

This paper describes the grid-oriented storage (GOS) architecture and its implementations. A GOS-specific file system (GOS-FS), the single-purpose intent of a GOS OS, and secure interfaces via grid security infrastructure (GSI) motivate and enable this new architecture. As an FTP server, GOS with a slimmed OS, with a total volume of around 150 MB, outperforms the standard GridFTP by 20-40 percent. As a file server, GOS-FS acts as a network/grid interface, enabling a user to perform searches and access resources without downloading them locally. In the real-world tests between Cambridge and Beijing, where the transfer distance is 10,000 km, the multistreamed GOS-FS file opening/saving resulted in a remarkable performance increase of about 2-25 times, compared to the single-streamed network file system (NFSv4). GOS is expected to be a variant of or successor to the well-used network-attached storage (NAS) and/or storage area network (SAN) products in the grid era

Dynamic and scalable storage management architecture for Grid Oriented Storage devices

Most of currently deployed Grid systems employ hierarchical or centralized approaches to simplify system management. However, the approaches cannot satisfy the requirements of complex Grid applications which involve hundreds or thousands of geographically distributed nodes. This paper proposes a Dynamic and Scalable Storage Management (DSSM) architecture for Grid Oriented Storage (GOS) devices. Since large-scale data intensive applications frequently involve a high degree of data access locality, the DSSM divides GOS nodes into multiple geographically distributed domains to facilitate the locality and simplify the intra-domain storage management. Dynamic GOS agents selected from the domains are organized as a virtual agent domain in a Peer-to-Peer (P2P) manner to coordinate multiple domains. As only the domain agents participate in the inter-domain communication, system wide information dissemination can be done far more efficiently than flat flooding. Grid service based storage resources are adopted to stack simple modular service piece by piece as demand grows. The decentralized architecture of DSSM avoids the hierarchical or centralized approaches of traditional Grid architectures, eliminates large-scale flat flooding of unstructured P2P systems, and provides an interoperable, seamless, and infinite storage pool in a Grid environment. The DSSM architecture is validated by a proof-of-concept prototype system.

Ad-hoc network of unmanned aerial vehicle swarms for search & destroy tasks

An ad-hoc network of unmanned aerial vehicles (UAVs) is modeled as a swarm of birds. Flock formation observed among birds in nature inspires the control of these UAVs, which perform a search and destroy task involving multiple, moving targets. The proposed control model is decentralized, adaptive and self-organizing to deal with the dynamic and distributed nature of the problem and relies only on local sensing and minimal communications to account for potential limitations in terms of global communications and lack of global information related to the task. The proposed model is tested for its self-organization capabilities in various simulation environments involving various number of UAVs and targets. These simulation show that the nature inspired control model is effective, robust and scalable in the context of the search and destroy tasks. Further simulations show that, because of the physical proximity of the UAVs within a swarm, a very good and robust routing performance can be achieved for these local, intra-swarm communications.

Cluster computing and grid 2005 works in progress

This is the second in a two-part series of works-in-progress articles taken from a special session, which was part of the Cluster Computing and Grid 2005 conference, held in Cardiff, UK. The session was organized by Mark Baker (University of Portsmouth, UK) and Daniel S. Katz (Jet Propulsion Laboratory, US). For more information, you can contact the session organizers or the authors of the articles.This is the second in a two-part series () of works-in-progress articles taken from a special session, which was part of the Cluster Computing and Grid 2005 conference (http://www.cs.cf.ac.uk/ccgrid2005), held in Cardiff, UK. The session was organized by Mark Baker (University of Portsmouth, UK) and Daniel S. Katz (Jet Propulsion Laboratory, US). For more information, you can contact the session organizers or the authors of the articles.

Evolutionary Storage: Speeding up a Magnetic Disk by Clustering Frequent Data

To counteract the performance degradation in magnetic disks, it is proposed to develop an evolutionary storage system that is evolving over time. The frequent data is clustered and moved to a flash memory. A large disk trace, collected from real systems, has been used. The experimental results showed that the performance speedup due to the disk reorganization is 4.0

Eight Times Acceleration of Geospatial Data Archiving and Distribution on the Grids

A grid-powered Web Geographical Information Science (GIS)/Web Processing Service (WPS) system has been developed for archiving and distributing large volumes of geospatial data. However, users, WPS servers, and data resources are always distributed across different locations, attempting to access and archive geospatial data from a GIS survey via conventional Hypertext Transport Protocol, Network File System Protocol, and File Transfer Protocol, which often encounters long waits and frustration in wide area network (WAN) environments. To provide a ldquolocal-likerdquo performance, a WAN/grid-optimized protocol known as ldquoGridJetrdquo developed at our lab was used as the underlying engine between WPS servers and clients, which utilizes a wide range of technologies including the one of paralleling the remote file access. No change in the way of using software is required since the multistreamed GridJet protocol remains fully compatible with the existing IP infrastructures. Our recent progress includes a real-world test that PyWPS and Google Earth over the GridJet protocol beat those over the classic ones by a factor of two to eight, where the distribution/archiving distance is over 10 000 km.

CloudJet4BigData: Streamlining Big Data via an Accelerated Socket Interface

Big data needs to feed users with fresh processing results and cloud platforms can be used to speed up big data applications. This paper describes a new data communication protocol (CloudJet) for long distance and large volume big data accessing operations to alleviate the large latencies encountered in sharing big data resources in the clouds. It encapsulates a dynamic multi-stream/multi-path engine at the socket level, which conforms to Portable Operating System Interface (POSIX) and thereby can accelerate any POSIX-compatible applications across IP based networks. It was demonstrated that CloudJet accelerates typical big data applications such as very large database (VLDB), data mining, media streaming and office applications by up to tenfold in real-world tests.

A Case for Redundant Arrays of Hybrid Disks (RAHD)

Hybrid Hard Disk Drive was originally concepted by Samsung, which incorporates a Flash memory in a magnetic disk. The combined ultra-high-density benefits of magnetic storage and the low-power and fast read access of NAND technology inspires us to construct Redundant Arrays of Hybrid Disks (RAHD) to offer a possible alternative to todays Redundant Arrays of Independent Disks (RAIDs) and/or Massive Arrays of Idle Disks (MAIDs). We first design an internal management system (including Energy-Efficient Control) for hybrid disks. Three traces collected from real systems as well as a synthetic trace are then used to evaluate the RAHD arrays. The trace-driven experimental results show: in the high speed mode, a RAHD outplays the purely-magnetic-disk-based RAIDs by a factor of 2.4-4; in the energy-efficient mode, a RAHD4/5 can save up to 89% of energy at little performance degradation.