Dániel Horváth

A Low Power, Programmable Networking Platform and Development Environment

Programmable networking platforms are getting widely used for customized traffic manipulation, analysis and network management. This propagates the need for exceptional development flexibility, for wide variety of high-speed interfaces and for the usage of high performance, yet low power technologies. This paper presents an FPGA-based programmable platform, capable of real-time processing, filtering and manipulating 10Gbps traffic. In order to expand its potential, besides the two 10GbE interfaces, the platform contains extension slots for COM express, mini PCI-e, and it has 16 onboard SFP connectors, towards which the fraction of the traffic, or even the full traffic can be forwarded to. The design is modular, programmable in both hardware (firmware) and software, aiming low power consumption. The full potential of the hardware can only be exploited with an easy-to-use development environment, with simple design customization and support for creating new applications. To fulfill this, a development environment is also presented, including a modeling framework that provides an easy way to create new networking applications on the platform. This framework allows modeling applications in SystemC, and eases the development of the hardware description code.

Joining of Tubular Parts by Electromagnetic Forming: Computational Investigations of Strength

A frequent application of electromagnetic forming in the industry is joining by electromagnetic compression of tubular parts. The goodness of the joints is determined by the strength of the joint, and it is considered excellent if it reaches the strength of the weaker material pair. Strength of these joint are generally tested for tensile and torsion loadings. In order to increase strength, one or more horizontal and/or radial grooves can be fabricated onto the male joining pair to utilize form fitting besides interference fit. In the present paper, the strength of the joint is studied by finite element simulations.

An energy-efficient FPGA-based packet processing framework

Modern packet processing hardware (e.g. IPv6-supported routers) demands high processing power, while it also should be power-efficient. In this paper we present an architecture for high-speed packet processing with a hierarchical chip-level power management that minimizes the energy consumption of the system. In particular, we present a modeling framework that provides an easy way to create new networking applications on an FPGA based board. The development environment consists of a modeling environment, where the new application is modeled in SystemC. Furthermore, our power management is modeled and tested against different traffic loads through extensive simulation analysis. Our results show that our proposed solution can help to reduce the energy consumption significantly in a wide range of traffic scenarios.

Ring Flushing for Reduced Overload in Spanning Tree Protocol Controlled Ethernet Networks

Flooding causes serious problems to the scalability of Ethernet networks. Recent proposals to overcome this problem, such as SEATTLE [5], usually require significant changes in different network layers, making the realistic chance of their deployment questionable. In this paper, we propose Ring Flushing, a practical method to reduce the burden of flooding during topology changes. The basic idea behind our approach is to locate stale forwarding information in an efficient way. Ring Flushing abolishes the broadcast-like spreading of topology change information thus shrinking the flushing domain. We implemented Ring Flushing in OMNeT++ simulation environment and evaluated its performance in different topologies and parameter settings. Our simulations show that the Ring Flushing has clear advantage over the approach of standard Rapid Spanning Tree Protocol (RSTP) in terms of throughput during network recovery. Furthermore, the Ring Flushing diminishes overall network overload during topology changes as the network size increases.

Shaping and reinforcement of melt textured YBa2Cu3O7−δ superconductors

From porous Y2BaCuO5 (Y211) with various grain sizes, single domain ceramic composites YBa2Cu3Oy/Y2BaCuO5 have been prepared by combination of the infiltration and top seed growth (ITSG) process. In addition, perforated Y123 has been prepared from Y211 by the ITSG method in order to magnify the specific surface and then increase oxygen diffusion into the core of the material. Magnetic and electrical properties were determined and correlate well with the microstructure of the composites and were compared to the conventional doped or undoped YBa2Cu3Oy (Y123). From magnetic measurements, high critical current densities, Jc, of 86 000 A cm−2 have been measured. Transport Jc values higher than 10 600 A cm−2 are reached at 77 K and 0 T, corresponding to the nominal critical currents of 325 A injected reproducibly through sections less than 3.082 mm2. This confirms the high quality of single domains obtained with a well controlled ITSG process. On the other hand, the perforated samples were reinforced using resin impregnation and the flux mapping has been investigated.

Joining of Tubular Parts by Electromagnetic Forming; Experimental Investigations

Electromagnetic forming is a high speed forming process, wherein the forming pressure is created by high energy density electromagnetic pulse. Besides direct shaping there are other application areas as well, so electromagnetic plastic forming is a potential field of creating joints between tube and rod-like components. Connecting components of dissimilar materials is an increasing demand in the manufacturing process of structures in the automotive industry. The application of new technologies, such as electrodynamic, especially electromagnetic forming, is a possible method to satisfy these demands. The article summarizes the most important fundamentals of electromagnetic forming; in particular, tube-rod joints, the main types of such joints; interference-fit and form-fit joints are described. Experiments, which were carried out producing tube-rod joints with electromagnetic forming, are also introduced. A new type of form-fit joints for tube-rod connections has been developed, which can withstand not only tensile loads but also torsion. Experiments and mechanical tests have proved the applicability of this kind of joints.

Aspects of Electrodynamic Forming Processes

Two types of electrodynamic forming process have been developed: electromagnetic and electrohydraulic forming. In the case of electromagnetic forming, the energy stored in a capacitor bank is discharged through a coil, which means that the electrical interaction between the coil and the plate or a tubular part to be formed results in deformation of the workpiece. However, in the case of electrohydraulic forming, the capacitor bank is discharged through a spark gap or filament wire; the deformation of the workpiece is due to the shockwaves, generated by the discharge process in a transmitting medium. In both processes, a large amount of energy is released in extremely short time, therefore these processes are considered to be high energy rate forming processes. These high energy rates, result in increasing the formability of the materials in many cases, and obtain significant deformations also for some materials that normally do not behave plastically. The utilization of the energy stored in the capacitor bank is significantly better in the case of electrohydraulic forming, because the released energy is converted directly to pressure waves, results in forming of higher strength materials. Both metallic and non-metallic materials can be formed by the technologies of electromagnetic and electrohydraulic technologies. In the present paper some aspects and applications of these high energy rate methods are briefly outlined mainly focusing on the automotive industry, involving expansion or compression forming of tubular parts, joining and assembly operations.

A SystemC-based simulation framework for energy-efficiency evaluation of embedded networking devices

In this paper, we discuss the design and the implementation details of a simulation framework which provides an easy-to-use environment for modeling of a network device. This framework comes with a packet processing architecture that can be modified and extended for a particular network device. Moreover, this framework enables the energy-consumption evaluation of the modeled network hardware by the application of energy management modules which can turn off unused modules. The framework is implemented using the SystemC programming language which is suitable for description of hardware and the software running on it. Finally, a use-case of 8-port gigabit switch based on the implemented packet processing framework is evaluated in different network scenarios.

Trapped field of YBCO single-domain samples using pulse magnetization from 77K to 20K

ReBCO single-domain bulk superconductors have been shown to trap significant magnetic field at 77K and below. They can advantageously replace permanent magnets in cryogenic motors; more power in a smaller volume can be achieved. But practically, their magnetization has to be performed in situ. Usually it implies the use of pulse magnetization which is severe for the samples. This technique generates heat and stress on the superconductors. The magnetic-flux-trapping capabilities of YBCO single-domain samples were explored using the pulse-field facilities at the LNCMP (National Pulsed Magnetic Field Laboratory) at Toulouse, France. The flux dynamic was monitored during magnetic pulses by measuring the surface induction with a Hall probe on top of the samples at different temperatures from 77K to 20K. The samples were 16 mm in diameter and about 10 mm in height. The best one trapped 400 mT at 77K and 2.5T at 20K. The trapped field increases almost linearly down to 40K. The magnetic pulse is seen to generate heat. The temperature rise increases with decreasing temperature dwell because of lower heat capacity. The achieved trapped field is a compromise between the temperature rise and the applied field, and depends greatly of the magnetization history.