Browsing by Subject "Flow control"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Flow control of real-time unicast multimedia applications in best-effort networks(2009-05-15) Bhattacharya, AnindaOne of the fastest growing segments of Internet applications are real-time mul- timedia applications, like Voice over Internet Protocol (VoIP). Real-time multimedia applications use the User Datagram Protocol (UDP) as the transport protocol because of the inherent conservative nature of the congestion avoidance schemes of Transmis- sion Control Protocol (TCP). The e?ects of uncontrolled ?ows on the Internet have not yet been felt because UDP tra?c frequently constitutes only ? 20% of the total Internet tra?c. It is pertinent that real-time multimedia applications become better citizens of the Internet, while at the same time deliver acceptable Quality of Service (QoS). Traditionally, packet losses and the increase in the end-to-end delay experienced by some of the packets characterizes congestion in the network. These two signals have been used to develop most known ?ow control schemes. The current research considers the ?ow accumulation in the network as the signal for use in ?ow control. The most signi?cant contribution of the current research is to propose novel end- to-end ?ow control schemes for unicast real-time multimedia ?ows transmitting over best-e?ort networks. These control schemes are based on predictive control of the accumulation signal. The end-to-end control schemes available in the literature are based on reactive control that do not take into account the feedback delay existing between the sender and the receiver nor the forward delay in the ?ow dynamics. The performance of the proposed control schemes has been evaluated using the ns-2 simulation environment. The research concludes that active control of hard real- time ?ows delivers the same or somewhat better QoS as High Bit Rate (HBR, no control), but with a lower average bit rate. Consequently, it helps reduce bandwidth use of controlled real-time ?ows by anywhere between 31:43% to 43:96%. Proposed reactive control schemes deliver good QoS. However, they do not scale up as well as the predictive control schemes. Proposed predictive control schemes are e?ective in delivering good quality QoS while using up less bandwidth than even the reactive con- trol schemes. They scale up well as more real-time multimedia ?ows start employing them.Item Flow control simulation with synthetic and pulsed jet actuator(2010-08) Jee, Sol Keun, 1979-; Moser, Robert deLancey; Bogard, David G.; Ezekoye, Ofodike A.; Goldstein, David B.; Raman, VenkatramananTwo active flow control methods are investigated numerically to understand the mechanism by which they control aerodynamics in the presence of severe flow separation on an airfoil. In particular, synthetic jets are applied to separated flows generated by additional surface feature (the actuators) near the trailing edge to obtain Coanda-like effects, and an impulse jet is used to control a stalled flow over an airfoil. A moving-grid scheme is developed, verified and validated to support simulations of external flow over moving bodies. Turbulent flow is modeled using detached eddy simulation (DES) turbulence models in the CFD code CDP (34) developed by Lopez (54). Synthetic jet actuation enhances turbulent mixing in flow separation regions, reduces the size of the separation, deflects stream lines closer to the surface and changes pressure distributions on the surface, all of which lead to bi-directional changes in the aerodynamic lift and moment. The external flow responds to actuation within about one convective time, which is significantly faster than for conventional control surfaces. Simulation of pitching airfoils shows that high-frequency synthetic jet affects the flow independently of the baseline frequencies associated with vortex shedding and airfoil dynamics. These unique features of synthetic jets are studied on a dynamically maneuvering airfoil with a closed-loop control system, which represents the response of the airfoil in wind-tunnel experiments and examines the controller for a rapidly maneuvering free-flight airfoil. An impulse jet, which is applied upstream of a nominal flow separation point, generates vortices that convect downstream, interact with the separating shear layer, dismantle the layer and allow following vortices to propagate along the surface in the separation region. These following vortices delay the separation point reattaching the boundary layer, which returns slowly to its initial stall condition, as observed in wind-tunnel experiments. A simple model of the impulse jet actuator used herein is found to be sufficient to represent the global effects of the jet on the stalled flow because it correctly represents the momentum injected into the flow.Item Investigation of a pulsed-plasma jet for separation shock/boundary layer interaction control(2010-05) Narayanaswamy, Venkateswa; Raja, Laxminarayan L.; Clemens, Noel T.; Varghese, Philip; Raman, Venkatramanan; Bengtson, RogerA pulsed-plasma jet (called a "spark-jet" by other researchers), is a high-speed synthetic jet that is generated by striking an electrical discharge in a small cavity. The gas in the cavity pressurizes owing to the heating and is allowed to escape through a small orifice. A series of experiments were conducted to determine the characteristics of the pulsed-plasma jet issuing into stagnant air at a pressure of 45 Torr. These results show that typical jet exit velocities of about 250 m/s can be induced with discharge energies of about 30 mJ per jet. Furthermore, the maximum pulsing frequency was found to be about 5 kHz, because above this frequency the jet begins to misfire. The misfiring appears to be due to the finite time it takes for the cavity to be recharged with ambient air between discharge pulses. The velocity at the exit of the jet is found to be primarily dependent on the discharge current and independent of other discharge parameters such as cavity volume and orifice diameter. Temperature measurements are made using optical emission spectroscopy and reveal the presence of considerable non-equilibrium between rotational and vibrational modes. The gas heating efficiency was found to be 10% and this parameter is shown to have a direct effect on the plasma jet velocity. These results indicate that the pulsed-plasma jet creates a sufficiently strong flow perturbation that is holds great promise as a supersonic flow actuator. An experimental study is conducted to characterize the performance of a pulsed-plasma jet for potential use in supersonic flow control applications. To obtain an estimate of the relative strength of the pulsed-plasma jet, the jet is injected normally into a Mach 3 cross-flow and the penetration distance is measured by using schlieren imaging. These measurements show that the jet penetrates 1.5 [delta], where [delta] is the boundary layer thickness, into the cross-flow and the jet-to-crossflow momentum flux ratio is estimated to be 0.6. An array of pulsed-plasma jets was issued from different locations upstream of a 30-degree compression ramp in a Mach 3 flow. Furthermore, two different jet configurations were used: normal injection and pitched and skewed injection. The pitched and skewed configuration was used to see if the jets could act as high-bandwidth pulsed vortex generators. The interaction between the jets and the separation shock was studied using phase-locked schlieren imaging. Results show that the plasma jets cause a significant disturbance to the separation shock and clearly influence its unsteadiness. While all plasma jet configurations tested caused an upstream motion of the separation shock, pitched and skewed plasma jets caused an initial downstream shock motion before the upstream motion, demonstrating the potential use of these plasma jets as vortex generator jets. The effect of the plasma jet array on the separation shock unsteadiness is studied in a time-resolved manner by using 10 kHz schlieren imaging and fast-response wall pressure measurements. An array of three pulsed-plasma jets, in a pitched and skewed configuration, is used to force the unsteady motion of the interaction formed by a 24° compression ramp in a Mach 3 flow. The Reynolds number of the incoming boundary layer is Re[theta]=3300. Results show that when the pulsed jet array is placed upstream of the interaction, the jets cause the separation shock to move in a quasi-periodic manner, i.e., nearly in sync with the pulsing cycle. As the jet fluid convects across the separation shock, the shock responds by moving upstream, which is primarily due to the presence of hot gas and hence the lower effective Mach number of the incoming flow. Once the hot gases pass through the interaction, the separation shock recovers by moving downstream, and this recovery velocity is approximately 1% to 3% of the free stream velocity. With forcing, the low-frequency energy content of the pressure fluctuations at a given location under the intermittent region decreases significantly. This is believed to be a result of an increase in the mean scale of the interaction under forced conditions. Pulsed-jet injection are also employed within the separation bubble, but negligible changes to the separation shock motion were observed. These results indicate that influencing the dynamics of this compression ramp interaction is much more effective by placing the actuator in the upstream boundary layer.Item Network-on-chip architectures for scalability and service guarantees(2011-08) Grot, Boris; Keckler, Stephen W.; Burger, Douglas C.; Mutlu, Onur; Witchel, Emmett; Zhang, YinRapidly increasing transistor densities have led to the emergence of richly-integrated substrates in the form of chip multiprocessors and systems-on-a-chip. These devices integrate a variety of discrete resources, such as processing cores and cache memories, on a single die with the degree of integration growing in accordance with Moore's law. In this dissertation, we address challenges of scalability and quality-of-service (QOS) in network architectures of highly-integrated chips. The proposed techniques address the principal sources of inefficiency in networks-on-chip (NOCs) in the form of performance, area, and energy overheads. We also present a comprehensive network architecture capable of interconnecting over a thousand discrete resources with high efficiency and strong guarantees. We first show that mesh networks, commonly employed in existing chips, fall significantly short of achieving their performance potential due to transient congestion effects that diminish network performance. Adaptive routing has the potential to improve performance through better load distribution. However, we find that existing approaches are myopic in that they only consider local congestion indicators and fail to take global network state into account. Our approach, called Regional Congestion Awareness (RCA), improves network visibility in adaptive routers via a light-weight mechanism for propagating and integrating congestion information. By leveraging both local and non-local congestion indicators, RCA improves network load balance and boosts throughput. Under a set of parallel workloads running on a 49-node substrate, RCA reduces on-chip network latency by 16%, on average, compared to a locally-adaptive router. Next, we target NOC latency and energy efficiency through a novel point-to-multipoint topology. Ring and mesh networks, favored in existing on-chip interconnects, often require packets to go through a number of intermediate routers between source and destination nodes, resulting in significant latency and energy overheads. Topologies that improve connectivity, such as fat tree and flattened butterfly, eliminate much of the router overhead, but require non-minimal channel lengths or large channel count, reducing energy-efficiency and/or performance as a result. We propose a new topology, called Multidrop Express Channels (MECS), that augments minimally-routed express channels with multi-drop capability. The resulting richly-connected NOC enjoys a low hop count with favorable delay and energy characteristics, while improving wire utilization over prior proposals. Applications such as virtualized servers-on-a-chip and real-time systems require chip-level quality-of-service (QOS) support to provide fairness, service differentiation, and guarantees. Existing network QOS approaches suffer from considerable performance and area overheads that limit their usefulness in a resource-limited on-die network. In this dissertation, we propose a new QOS scheme called Preemptive Virtual Clock (PVC). PVC uses a preemptive approach to provide hard guarantees and strong performance isolation while dramatically reducing queuing requirements that burden prior proposals. Finally, we introduce a comprehensive network architecture that overcomes the bottlenecks of earlier designs with respect to area, energy, and QOS in future highly-integrated chips. The proposed NOC uses a topology-centric QOS approach that restricts the extent of hardware QOS support to a fraction of the network without compromising guarantees. In doing so, network area and energy efficiency are significantly improved. Further improvements are derived through a novel flow-control mechanism, along with switch- and link-level optimizations. In concert, these techniques yield a network capable of interconnecting over a thousand terminals on a die while consuming 47% less area and 26% less power than a state-of-the-art QOS-enabled NOC. The mechanisms proposed in this dissertation are synergistic and enable efficient, high-performance interconnects for future chips integrating hundreds or thousands of on-die resources. They address deficiencies in routing, topologies, and flow control of existing architectures with respect to area, energy, and performance scalability. They also serve as a building block for cost-effective advanced services, such as QOS guarantees at the die level.