Cluster scheduling and load balancing via TCP options

This paper describes an experiment in designing, implementing and testing a Transport layer cluster scheduling and dispatching architecture. The motivation for the experiment was the hypothesis that a Transport layer clustering solution may offer advantantages over the existing industry-standard Network layer and Data Link Layer approaches. The critical success factors initially established to guide and evaluate the experiment were reduced dispatcher work load, reduced dispatcher internal state memory requirements, distributed denial of service resilience, and cluster software design simplicity. The functional design stage of the experiment produced a Transport layer strategy for scheduling and load balancing based on the specification of two new TCP options. Implementation required the introduction of the newly specified TCP options into the Linux (2.4) kernel. The implementation produced an extended Linux Socket API to facilitate user-process access to the additional TCP capability. The testing stage of the experiment confirmed the operational efficiency of the solution.

Examining TCP parallelization related methods for various packet losses

The diversity of the networks (wired/wireless) prefers a TCP solution robust across a wide range of networks rather than fine-tuned for a particular one at the cost of another. TCP parallelization uses multiple virtual TCP connections to transfer data for an application process and opens a way to improve TCP performance across a wide range of environments - high bandwidth-delay product (BDP), wireless as well as conventional networks. In particular, it can significantly benefit the emerging high-speed wireless networks. Despite its potential to work well over a wide range of networks, it is not fully understood how TCP parallelization performs when experiencing various packet losses in the heterogeneous environment. This paper examines the current TCP parallelization related methods under various packet losses and shows how to improve the performance of TCP parallelization.

Punching shear of slab-column connection in flat plate construction

The flat plate system is currently widely used in construction. It permits architectural flexibility, more clear space, less building height, easier formwork, and shorter construction time. However, there remains the problem of brittle punching failure due to the transfer of shearing forces and unbalanced moments at the flat plate-column connection. It is the purpose of this paper to investigate the effects of various interdependent factors that govern the punching shear resistance and behaviour of the flat plate-column connection, as well as their inclusion in current Codes.

PhotoVoltaic DC-DC module integrated converter for novel cascaded and bypass grid connection topologies - design and optimisation

Grid connected PhotoVoltaic (PV) inverters fall into three broad categories — Central, String and Module Integrated Converers (MICs). MICs offer any avantaes in performance and flexibility, but are at a cost disadvantage. Two alternative novel approaches proposed by the author — cascaded dc-dc MICs and bypass dc-dc MICs — integrate a simple non-isolated intelligent dc-dc converter with each PV module to provide the advantages of dc-ac MICs at a lower cost. A suitable universal 150W 5A dc-dc converter design is presented based on two interleaved MOSFET half bridges. Testing shows Zero Voltage Switching (ZVS) keeps losses under 1W for bi-directional power flows up to 15W between two adjacent 12V PV modules for the bypass application, and efficiencies over 94% for most of the operational power range for the cascaded converter application. Based on the experimental results, potential optimizations to further reduce losses are discussed.

Cascaded DC-DC converter connection of photovoltaic modules

New residential scale photovoltaic (PV) arrays are commonly connected to the grid by a single DC-AC inverter connected to a series string of PV modules, or many small DC-AC inverters which connect one or two modules directly to the AC grid. This paper shows that a "converter-per-module" approach offers many advantages including individual module maximum power point tracking, which gives great flexibility in module layout, replacement, and insensitivity to shading; better protection of PV sources, and redundancy in the case of source or converter failure; easier and safer installation and maintenance; and better data gathering. Simple nonisolated per-module DC-DC converters can be series connected to create a high voltage string connected to a simplified DC-AC inverter. These advantages are available without the cost or efficiency penalties of individual DC-AC grid connected inverters. Buck, boost, buck-boost and Cuk converters are possible cascadable converters. The boost converter is best if a significant step up is required, such as with a short string of 12 PV modules. A string of buck converters requires many more modules, but can always deliver any combination of module power. The buck converter is the most efficient topology for a given cost. While flexible in voltage ranges, buck-boost and Cuk converters are always at an efficiency or alternatively cost disadvantage.