Browsing by Subject "networking"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item A Multi-FPGA Networking Architecture and Its Implementation(2015-05-12) Knezek, Gabriel SFPGAs show great promise in accelerating compute-bound parallelizable applications by offloading kernels into programmable logic. However, currently FPGAs present significant hurdles in being a viable technology, due to both the capital outlay required for specialized hardware as well as the logic required to support the offloaded kernels on the FPGA. This thesis seeks to change that by making it easy to communicate clusters of FPGAs over IP networks and providing infrastructure for common application use cases, allowing authors to focus on their application and not the procurement and details of interacting with a specific FPGA. Our approach is twofold. First, we develop an FPGA IP network stack and bitfile management system allowing users to upload their logic to a server and have it run on FPGAs accessible through the Internet. Second, we engineer a programmable logic interface which authors can use to move data to their application kernels. This interface provides communication over the Internet as well as the scaffolding typically re-invented for each application by providing I/O between application logic, even if spread across different FPGAs. We utilize Partial Reconfiguration to divide the FPGAs into regions, each of which can host different applications from different users. We then provide a web service through which users can upload their FPGA logic. The service finds a spot for the logic on the FPGAs, reconfigures them to contain the logic, then sends back the user their IP addresses. To ease development of the application pieces themselves, our framework abstracts away the complexity of communicating over IP networks as well as between different FPGAs. Instead we provide an interface to applications consisting simply of a RAM port. Applications write packets of data into the port, and they appear at the other end, whether that other end is across an IP network or another FPGA. Finally, we then prove the feasibility and utility of our approach by implementing it on an array of Xilinx Virtex 5 FPGAs, linked together with GTP serial links and connected via Gigabit Ethernet. We port a compute-bound application based on regular expression string matching to the framework, demonstrating that our approach is feasible for implementing a realistic application.Item Low complexity channel models for approximating flat Rayleigh fading in network simulations(Texas A&M University, 2004-09-30) McDougall, Jeffrey MichaelThe intricate dependency of networking protocols upon the performance of the wireless channel motivates the investigation of network channel approximations for fading channels. Wireless networking protocols are increasingly being designed and evaluated with the assistance of networking simulators. While evaluating networking protocols such as medium access control, routing, and reliable transport, the network channel model, and its associated capacity, will drastically impact the achievable network throughput. Researcher relying upon simulation results must therefore use extreme caution to ensure the use of similar channel models when performing protocol comparisons. Some channel approximations have been created to mimic the behavior of a fading environment, however there exists little to no justification for these channel approximations. This dissertation addresses the need for a computationally efficient fading channel approximation for use in network simulations. A rigorous flat fading channel model was developed for use in accuracy measurements of channel approximations. The popular two-state Markov model channel approximation is analyzed and shown to perform poorly for low to moderate signal-to-noise ratios (SNR). Three novel channel approximations are derived, with multiple methods of parameter estimation. Each model is analyzed for both statistical performance and network performance. The final model is shown to achieve very accurate network throughput performance by achieving a very close matching of the frame run distributions. This work provides a rigorous evaluation of the popular two-state Markov model, and three novel low complexity channel models in both statistical accuracy and network throughput performance. The novel models are formed through attempts to match key statistical parameters of frame error run and good frame run statistics. It is shown that only matching key parameters is insufficient to achieve an acceptable channel approximation and that it is necessary to approximate the distribution of frame error duration and good frame run duration. The final novel channel approximation, the three-state run-length model, is shown to achieve a good approximation of the desired distributions when some key statistical parameters are matched.