Gigabit Ethernet Network Research Articles

Results from the first p+p runs of the ALICE High Level Trigger at LHC

The High Level Trigger for the ALICE experiment at LHC is a powerful, sophisticated tool aimed at compressing the raw data volume and issuing selective triggers for events with desirable physics content. At its current state it integrates information from all major ALICE detectors, i. e. the inner tracking system, the time projection chamber, the electromagnetic calorimeters, the transition radiation detector and the muon spectrometer performing real-time event reconstruction.The steam engine behind HLT is a high performance computing cluster of several hundred nodes. It has to reduce the data rate from 25 GB/s to 1.25 GB/s for fitting the DAQ mass storage bandwidth. The cluster is served by a full GigaBit Ethernet network, in addition to an InfiniBand backbone network. To cope with the great challenge of Pb+Pb collisions in autumn 2010, its performance capabilities are being enhanced with the addition of new nodes. Towards the same end the first GPU co-processors are in place.During the first period of data taking with p+p collisions the HLT was extensively used to reconstruct, analyze and display data from the various participating detectors. Among other tasks it contributed to the monitoring of the detector performance, selected events for their calibration and efficiency studies, and estimated primary and secondary vertices from p+p collisions identifying V0 topologies. The experience gained during these first months of online operation will be presented.

Compact TO-CAN Header With Bandwidth Excess 40 GHz

A large bandwidth transistor outline (TO)-CAN header used in a low cost laser module was designed through a three dimensional full wave electromagnetic simulation and demonstrated by a direct measurement technique. The simulation results show that the proposed TO-CAN header provides a resonance-free transmission bandwidth in excess of 40 GHz with a reflection loss of 25 GHz under -10 dB with an ideal 50 Ω terminal. A pair of specially designed K type adaptors was used to accommodate the TO-CAN header for the high frequency measurements. This is a direct contact setup, without extra parasitic effects, and allowing a near resonance-free measurement. The experimental results are in good agreement with simulated results in the frequency range up to 40 GHz. A time-domain reflectometry measurement and a 40 Gbit/s eye diagram test were made to verify the performance. In addition, a laser module packaged with the TO-CAN header is proposed. By taking the bonding wires into account, the simulated transmission bandwidth of the module is over 35 GHz and a clear 25 Gbit/s eye diagram is obtained. This cost effective package solution of laser module can be applied in the next generation 100 Gigabit Ethernet (100 GbE) network.

The LHCb Eventbuilder: Design, Implementation and Operational Experience

The LHCb event-builder is implemented using a large Gigabit Ethernet network using a push-protocol for a single stage read-out at a 1 MHz event injection rate. The destination assignment and dynamic load-balancing are facilitated by LHCb's Timing and Fast Control system. The assembly of event fragments is done on each event-filter farm node instead of having dedicated builder units. The design of the event-builder will be shortly described, followed by a description of the implementation, the protocol used and the performance during first data taking. The emphasis will be on the experience we gained during the running of such a large event-building network. We will discuss the problems we encountered and how we overcame them.

ATLAS-detector operations with beams: the data-acquisition perspective

The ATLAS data acquisition system is based on about 2k processing nodes, interconnected by a multi-layer Gigabit Ethernet network. About 20k applications provide the needed capabilities in terms of run control, event selection, data flow, local storage and data monitoring. The whole system has been successfully commissioned during the last two years with cosmic ray and calibration data and it turned out to be robust and reliable. Nevertheless, the continuous operation with beams, the concurrent trigger commissioning, and the understanding of detector and physics performance pose new challenges. The flexibility of the data acquisition infrastructure will be probed and exploited, in order to comply with unseen working conditions in terms of data flow, monitoring and configuration requirements. The goal is to minimise both downtime and dead-time, allowing for runtime reconfiguration of the data acquisition and sub-detectors systems as well as for automatic error handling and recovery.

Error Detection and Reconfigurationin Reliable Ethernet Train Networks

In this paper, a novel reconfiguration technique is developed in the context of a fault-tolerant Networked Control System (NCS) in two train wagons. All sensors, controllers and actuators in both wagons are connected on top of a single Gigabit Ethernet network. The network also carries wired and wireless entertainment loads. A Markov model is used to prove that this reconfiguration technique reduces the effect of a failure in the error detection and switching mechanisms on the reliability of the control function. All calculations are based on closed-form solutions and verified using the SHARPE software package.

Design and Simulation of an Optical Gigabit Ethernet Network

This paper deals with the design and simulation of an optical gigabit Ethernet using OPTSIM 3.6 packet software. The main aim of the proposed design is to build a MAN optical network using one-gigabit Ethernet technique, and what are the necessary requirements to build these networks. As a case study, all states center are connected as Star – Bus topology using layer2 and layer3 optical switches. In addition, in this paper one-gigabit optical transmitter and receiver are designed to work as a node in the network topology. Further more, the benefits of using L- Band wavelength for transmission take in consider the linear and non-linear effects on fiber optic is presented.

Towards performance portability through runtime adaptation for high‐performance computing applications

AbstractThe Abstract Data and Communication Library (ADCL) is an adaptive communication library optimizing application level group communication operations at runtime. The library provides for a given communication pattern a large number of implementations and incorporates a runtime selection logic in order to choose the implementation leading to the highest performance. In this paper, we demonstrate how an application utilizing ADCL is deployed on a wide range of HPC architectures, including an IBM Blue Gene/L, an NEC SX‐8, an IBM Power PC cluster using an IBM Federation Switch, an AMD Opteron cluster utilizing a 4xInfiniBand and a Gigabit Ethernet network, and an Intel EM64T cluster using a hierarchical Gigabit Ethernet network with reduced uplink bandwidth. We demonstrate, how different implementations for the three‐dimensional neighborhood communication lead to the minimal execution time of the application on different architectures. ADCL gives the user the advantage of having to maintain only a single version of the source code and still have the ability to achieve close to optimal performance for the application on all architectures. Copyright © 2010 John Wiley & Sons, Ltd.

ATLAS DataFlow Infrastructure: Recent results from ATLAS cosmic and first-beam data-taking

The ATLAS DataFlow infrastructure is responsible for the collection and conveyance of event data from the detector front-end electronics to the mass storage. Several optimized and multi-threaded applications fulfill this purpose operating over a multi-stage Gigabit Ethernet network which is the backbone of the ATLAS Trigger and Data Acquisition System. The system must be able to efficiently transport event-data with high reliability, while providing aggregated bandwidths larger than 5 GByte/s and coping with many thousands network connections. Nevertheless, routing and streaming capabilities and monitoring and data accounting functionalities are also fundamental requirements.During 2008, a few months of ATLAS cosmic data-taking and the first experience with the LHC beams provided an unprecedented test-bed for the evaluation of the performance of the ATLAS DataFlow, in terms of functionality, robustness and stability. Besides, operating the system far from its design specifications helped in exercising its flexibility and contributed in understanding its limitations. Moreover, the integration with the detector and the interfacing with the off-line data processing and management have been able to take advantage of this extended data taking-period as well.In this paper we report on the usage of the DataFlow infrastructure during the ATLAS data-taking. These results, backed-up by complementary performance tests, validate the architecture of the ATLAS DataFlow and prove that the system is robust, flexible and scalable enough to cope with the final requirements of the ATLAS experiment.

Architecture and Implementation of the Front-End Electronics of the Time Projection Chambers in the T2K Experiment

The tracker of the near detector in the T2K neutrino oscillation experiment comprises three time projection chambers based on micro-pattern gaseous detectors. A new readout system is being developed to amplify, condition and acquire in real time the data produced by the 124.000 detector channels. The cornerstone of the system is a 72-channel application specific integrated circuit which is based on a switched capacitor array. Using analog memories combined with deferred digitization enables reducing the initial burstiness of traffic from 50 Tbps to 400 Gbps in a practical manner and with a very low power budget. Modern field programmable gate arrays coupled to commercial digital memories are the next elements in the chain. Multi-gigabit optical links provide 140 Gbps of aggregate bandwidth to carry data outside of the magnet surrounding the detector to concentrator cards that pack data and provide the interface to commercial PCs via a standard Gigabit Ethernet network. We describe the requirements and constraints for this application and justify our technical choices. We detail the design and the performance of several key elements and show the deployment of the front-end electronics on the first time projection chamber where the final tests before installation on-site are being conducted.

Quality of service control in Ethernet passive optical networks based on virtual private LAN service technique

A technique is proposed that allows achievement of logical point-to-multipoint paths in gigabit Ethernet (GbE) networks with access based on passive optical networks (PONs), guaranteeing upstream/downstream bandwidth and quality of service also in conditions of traffic congestion. Such a method is based on the combination of the virtual private LAN service (VPLS), implemented in the core network, and the VLAN tagging forwarding process used in the edge-PON segment.

RADIOSTAR: Providing wireless coverage over Gigabit Ethernet

RadioStar is a novel system developed by Bell Labs to provide low cost, cellular wireless coverage primarily targeted for enterprises, public spaces, and residences. The underlying concept is to digitize the output radio signals of a base station and to transport these signals over a Gigabit Ethernet network to multiple low-cost radio frequency radiation points. The received signals are returned through the network to the base station. This paper presents the architecture, implementation, and demonstration of RadioStar. It is shown how handoffs and precise frequency synchronization and timing control over a Gigabit Ethernet network can be established in such a way that the most stringent requirements of cellular communication air interfaces are satisfied. By converging cellular signal transport onto Gigabit Ethernet data networks, RadioStar can achieve substantially lower hardware and installation costs, while providing higher performance, more flexibility, and better operation, administration, and maintenance than current solutions. RadioStar concepts are being used in the ClearFill Star in-building product developed by the RFS business division of Alcatel-Lucent and currently being deployed commercially.

Comparative packet-forwarding measurement of three popular operating systems

This papers measures and compares the network performance (with respect to packet forwarding) of three popular operating systems when used in today's Gigabit Ethernet networks. Specifically, the paper compares the performance in terms of packet forwarding of Linux, Windows Server and Windows XP. We measure both kernel- and user-level packet forwarding when subjecting hosts to different traffic load conditions. The performance is compared and analyzed in terms of throughput, packet loss, delay, and CPU availability. Our evaluation methodology is based on packet-forwarding measurement which is a standard and popular benchmark and evaluation methodology to assess the performance of network elements such as servers, gateways, routers, and switches. Our evaluation methodology considers different configuration setups and utilizes open-source software tools to generate relatively high traffic rates. We consider today's typical network hosts of modern processors and Gigabit network cards. Our measurements show that in general Linux exhibits superior overall performance in the case of kernel (or IP) packet forwarding, whereas Windows Server exhibits superior performance in the case of user-level packet forwarding.

Design and implementation of a protocol offload engine for TCP/IP and remote direct memory access based on hardware/software coprocessing

This paper presents the design and implementation of a protocol offload engine that processes TCP/IP and remote direct memory access (RDMA) protocols by means of hardware/software coprocessing. In the offload engine, time-consuming operations such as TCP/IP header generation are implemented as hardware to improve performance. The software performs control operations and RDMA header generation. In the experiments and analyses, it is proved that the hardware can provide satisfactory performance to process all operations at speeds of over 1 Gbps. Our engine can offload most protocol processing overheads – up to 95% to 100% – from the host CPU. Finally, although the embedded processors operate with a 300 MHz clock that is seven times slower than the clock of the host CPU, our engine shows maximum bandwidths of 673 Mbps for TCP/IP and 551 Mbps for RDMA on a gigabit Ethernet network.

Performance Comparison of IP-Networked Storage

Dramatically increasing amounts of digital data are placing huge requirements on storage systems. IP-networked storage systems, such as the network file system (NFS)-based network-attached storage (NAS) systems and the iSCSI-storage area network (SAN) systems, have become increasingly common in today's local area network (LAN) environments. The emergence of new storage techniques, such as object-based storage (OBS) and content aware storage (CAS), significantly improves the functionality of storage devices to meet further needs for storage sub-systems. However, these may impact system performance. This paper compares the performance of NFS, iSCSI storage, object-based storage devices (OSDs), and CAS-based storage systems in an environment with no data sharing across host machines. A gigabit ethernet network is used as the storage network. Test results demonstrate that the performances of these systems are comparable with CAS being much better than the others for write operations. The performance bottlenecks in these systems are analyzed to provide insight into how future storage systems may be improved and possible optimization methods. The analysis shows how the I/O interfaces in these systems affect the application performance and that network-based storage systems require optimized I/O latency and reduced network and buffer processing in the servers.

Real-time distributed vision-based network system for logistics applications

The communication network providing real-time image sharing between the planning office and the warehouse for supply chain management is presented in this paper. Middleware plays a major role in developing distributed system efficiently for communication between image capturing devices (Vision Node) and image processing node as well as the distributed nodes. The image distribution over the network in real-time manner is useful to numerous logistic applications, such as quality control of pick-and-pack service, planning of cargo and pallet loading, and vehicles tracking. A detail construction of the vision-based network system both with respect to image capturing devices and data transmission has been announced. A gigabit Ethernet network is acting as a backbone of the whole system. The employment of the Network Data Distribution Service (NDDS) for the real-time distributed vision-based network system is discussed and experimental result of the real-time and delay time properties in each part on the network were considered and presented in this paper. Test results show that NDDS is capable of supporting a distributed logistics vision system.

High-performance 850 nm vertical-cavity surface-emitting laser in Gigabit Ethernet network

High-performance oxide vertical-cavity surface-emitting (VCSEL) laser is fabricated, and its usefulness is demonstrated as a suitable transmitting light source at 850 nm operating wavelength for Gigabit Ethernet application. Utilization of barrier reduction layers reveals low-threshold current requirement for operation at high modulation bandwidth. The electrical and optical characteristics, measured from the fabricated VCSEL, are simulated for Gigabit Ethernet transmission. Data rates of 1.25 Gbps with a bit error rate of 10 −11 are achieved by the use of a specific multimode network simulator.

LHCb Online event processing and filtering

The first level trigger of LHCb accepts one million events per second. After preprocessing in custom FPGA-based boards these events are distributed to a large farm of PC-servers using a high-speed Gigabit Ethernet network. Synchronisation and event management is achieved by the Timing and Trigger system of LHCb. Due to the complex nature of the selection of B-events, which are the main interest of LHCb, a full event-readout is required. Event processing on the servers is parallelised on an event basis. The reduction factor is typically 1/500. The remaining events are forwarded to a formatting layer, where the raw data files are formed and temporarily stored. A small part of the events is also forwarded to a dedicated farm for calibration and monitoring. The files are subsequently shipped to the CERN Tier0 facility for permanent storage and from there to the various Tier1 sites for reconstruction. In parallel files are used by various monitoring and calibration processes running within the LHCb Online system. The entire data-flow is controlled and configured by means of a SCADA system and several databases. After an overview of the LHCb data acquisition and its design principles this paper will emphasize the LHCb event filter system, which is now implemented using the final hardware and will be ready for data-taking for the LHC startup. Control, configuration and security aspects will also be discussed.

Performance analysis of parallel FDTD method on the different platforms

AbstractIn this communication, we investigate the performance of parallel FDTD code on a Gigabit Ethernet and optical network. It is fact that the optical network is necessary for a large cluster due to its wide bandwidth and small latency. However, for a small cluster, the Gigabit Ethernet will have a better ratio of performance to price. We also investigate the difference between high performance server and regular PC for the parallel FDTD code. © 2008 Wiley Periodicals, Inc. Microwave Opt Technol Lett 50: 2465–2467, 2008; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.23661

Experimental investigation on optical gigabit Ethernet network reliability for high-definition IPTV services

We show an experimental investigation about the quality of service of high-definition IPTV applications in a core-access gigabit Ethernet-multiprotocol label switching (MPLS) network test bed in terms of objective (or network) and subjective (or perceptive) measurements, using different network configurations. In the core part we investigated the role of the classes of services defined according to the differentiated service over MPLS and to the virtual private LAN service approach. Furthermore we also investigated the role of the restoration consequent to a link failure. For the access part we will show that for the asymmetric digital subscriber line 2+ (ADSL2+) architecture an important role is given to the edge routers to select the traffic with higher priority. Furthermore we show how an upgrade of the passive optical networks, by means of the WDM technique, can be particularly suitable for future peer-to-peer huge bandwidth services.

The Development of the New Data Acquisition System Without Hardware Trigger for the Super-Kamiokande Experiment

An online system of the underground neutrino detector, Super-Kamiokande is scheduled to be upgraded in 2008 together with front-end electronics. This detector consists of 50 000 tons of pure water equipped with about 13 000 photo-multipliers (PMTs) to detect Cherenkov light. The new online system is required to accept the dataflow of up to 800 MB/s from the front-end electronics and process them for the offline analysis. We will utilize a Gigabit Ethernet network and parallel data processing to handle this large amount of flow. In the new data acquisition scheme, we will not use a hardware event-trigger but read out every hit data from the front-end electronics and process them by the online farm. Therefore, there is no threshold of the number of PMT hits and the detector will become more sensitive to important cosmic-ray events such as relic neutrino from supernova and low energy solar neutrino. In addition to that, a dead-timeless system is desirable for the continuous measurement 365 days a year. In this paper, the detailed design of the upgraded online system and testing activities using prototypes will be described.