Host and network optimizations for performance enhancement and energy efficiency in data center networks

Modern data centers host hundreds of thousands of servers to achieve economies of scale. Such a huge number of servers create challenges for the data center network (DCN) to provide proportionally large bandwidth. In addition, the deployment of virtual machines (VMs) in data centers raises the requirements for efficient resource allocation and find-grained resource sharing. Further, the large number of servers and switches in the data center consume significant amounts of energy. Even though servers become more energy efficient with various energy saving techniques, DCN still accounts for 20% to 50% of the energy consumed by the entire data center. The objective of this dissertation is to enhance DCN performance as well as its energy efficiency by conducting optimizations on both host and network sides. First, as the DCN demands huge bisection bandwidth to interconnect all the servers, we propose a parallel packet switch (PPS) architecture that directly processes variable length packets without segmentation-and-reassembly (SAR). The proposed PPS achieves large bandwidth by combining switching capacities of multiple fabrics, and it further improves the switch throughput by avoiding padding bits in SAR. Second, since certain resource demands of the VM are bursty and demonstrate stochastic nature, to satisfy both deterministic and stochastic demands in VM placement, we propose the Max-Min Multidimensional Stochastic Bin Packing (M3SBP) algorithm. M3SBP calculates an equivalent deterministic value for the stochastic demands, and maximizes the minimum resource utilization ratio of each server. Third, to provide necessary traffic isolation for VMs that share the same physical network adapter, we propose the Flow-level Bandwidth Provisioning (FBP) algorithm. By reducing the flow scheduling problem to multiple stages of packet queuing problems, FBP guarantees the provisioned bandwidth and delay performance for each flow. Finally, while DCNs are typically provisioned with full bisection bandwidth, DCN traffic demonstrates fluctuating patterns, we propose a joint host-network optimization scheme to enhance the energy efficiency of DCNs during off-peak traffic hours. The proposed scheme utilizes a unified representation method that converts the VM placement problem to a routing problem and employs depth-first and best-fit search to find efficient paths for flows.

Energy-efficient self -organization of wireless acoustic sensor networks for ground target tracking

With the developments in computing and communication technologies, wireless sensor networks have become popular in wide range of application areas such as health, military, environment and habitant monitoring. Moreover, wireless acoustic sensor networks have been widely used for target tracking applications due to their passive nature, reliability and low cost. Traditionally, acoustic sensor arrays built in linear, circular or other regular shapes are used for tracking acoustic sources. The maintaining of relative geometry of the acoustic sensors in the array is vital for accurate target tracking, which greatly reduces the flexibility of the sensor network. To overcome this limitation, we propose using only a single acoustic sensor at each sensor node. This design greatly improves the flexibility of the sensor network and makes it possible to deploy the sensor network in remote or hostile regions through air-drop or other stealth approaches. Acoustic arrays are capable of performing the target localization or generating the bearing estimations on their own. However, with only a single acoustic sensor, the sensor nodes will not be able to generate such measurements. Thus, self-organization of sensor nodes into virtual arrays to perform the target localization is essential. We developed an energy-efficient and distributed self-organization algorithm for target tracking using wireless acoustic sensor networks. The major error sources of the localization process were studied, and an energy-aware node selection criterion was developed to minimize the target localization errors. Using this node selection criterion, the self-organization algorithm selects a near-optimal localization sensor group to minimize the target tracking errors. In addition, a message passing protocol was developed to implement the self-organization algorithm in a distributed manner. In order to achieve extended sensor network lifetime, energy conservation was incorporated into the self-organization algorithm by incorporating a sleep-wakeup management mechanism with a novel cross layer adaptive wakeup probability adjustment scheme. The simulation results confirm that the developed self-organization algorithm provides satisfactory target tracking performance. Moreover, the energy saving analysis confirms the effectiveness of the cross layer power management scheme in achieving extended sensor network lifetime without degrading the target tracking performance.

Proton Form Factor Puzzle and the CEBAF Large Acceptance Spectrometer (CLAS) two-photon exchange experiment

The electromagnetic form factors are the most fundamental observables that encode information about the internal structure of the nucleon. The electric (GE) and the magnetic ( GM) form factors contain information about the spatial distribution of the charge and magnetization inside the nucleon. A significant discrepancy exists between the Rosenbluth and the polarization transfer measurements of the electromagnetic form factors of the proton. One possible explanation for the discrepancy is the contributions of two-photon exchange (TPE) effects. Theoretical calculations estimating the magnitude of the TPE effect are highly model dependent, and limited experimental evidence for such effects exists. Experimentally, the TPE effect can be measured by comparing the ratio of positron-proton elastic scattering cross section to that of the electron-proton [R = σ(e +p)/σ(e+p)]. The ratio R was measured over a wide range of kinematics, utilizing a 5.6 GeV primary electron beam produced by the Continuous Electron Beam Accelerator Facility (CEBAF) at Jefferson Lab. This dissertation explored dependence of R on kinematic variables such as squared four-momentum transfer (Q2) and the virtual photon polarization parameter (&epsis;). A mixed electron-positron beam was produced from the primary electron beam in experimental Hall B. The mixed beam was scattered from a liquid hydrogen (LH2) target. Both the scattered lepton and the recoil proton were detected by the CEBAF Large Acceptance Spectrometer (CLAS). The elastic events were then identified by using elastic scattering kinematics. This work extracted the Q2 dependence of R at high &epsis;(&epsis; > 0.8) and the $&epsis; dependence of R at ⟨Q 2⟩ approx 0.85 GeV2. In these kinematics, our data confirm the validity of the hadronic calculations of the TPE effect by Blunden, Melnitchouk, and Tjon. This hadronic TPE effect, with additional corrections contributed by higher excitations of the intermediate state nucleon, largely reconciles the Rosenbluth and the polarization transfer measurements of the electromagnetic form factors.

Effective energy conservation and management in the building sector : the answer to the energy predicament

Modern civilization has developed principally through man's harnessing of forces. For centuries man had to rely on wind, water and animal force as principal sources of power. The advent of the industrial revolution, electrification and the development of new technologies led to the application of wood, coal, gas, petroleum, and uranium to fuel new industries, produce goods and means of transportation, and generate the electrical energy which has become such an integral part of our lives. The geometric growth in energy consumption, coupled with the world's unrestricted growth in population, has caused a disproportionate use of these limited natural resources. The resulting energy predicament could have serious consequences within the next half century unless we commit ourselves to the philosophy of effective energy conservation and management. National legislation, along with the initiative of private industry and growing interest in the private sector has played a major role in stimulating the adoption of energy-conserving laws, technologies, measures, and practices. It is a matter of serious concern in the United States, where ninety-five percent of the commercial and industrial facilities which will be standing in the year 2000 - many in need of retrofit - are currently in place. To conserve energy, it is crucial to first understand how a facility consumes energy, how its users' needs are met, and how all internal and external elements interrelate. To this purpose, the major thrust of this report will be to emphasize the need to develop an energy conservation plan that incorporates energy auditing and surveying techniques. Numerous energy-saving measures and practices will be presented ranging from simple no-cost opportunities to capital intensive investments.

Energy-Efficient Self-Organization of Wireless Acoustic Sensor Networks for Ground Target Tracking

With the developments in computing and communication technologies, wireless sensor networks have become popular in wide range of application areas such as health, military, environment and habitant monitoring. Moreover, wireless acoustic sensor networks have been widely used for target tracking applications due to their passive nature, reliability and low cost. Traditionally, acoustic sensor arrays built in linear, circular or other regular shapes are used for tracking acoustic sources. The maintaining of relative geometry of the acoustic sensors in the array is vital for accurate target tracking, which greatly reduces the flexibility of the sensor network. To overcome this limitation, we propose using only a single acoustic sensor at each sensor node. This design greatly improves the flexibility of the sensor network and makes it possible to deploy the sensor network in remote or hostile regions through air-drop or other stealth approaches. Acoustic arrays are capable of performing the target localization or generating the bearing estimations on their own. However, with only a single acoustic sensor, the sensor nodes will not be able to generate such measurements. Thus, self-organization of sensor nodes into virtual arrays to perform the target localization is essential. We developed an energy-efficient and distributed self-organization algorithm for target tracking using wireless acoustic sensor networks. The major error sources of the localization process were studied, and an energy-aware node selection criterion was developed to minimize the target localization errors. Using this node selection criterion, the self-organization algorithm selects a near-optimal localization sensor group to minimize the target tracking errors. In addition, a message passing protocol was developed to implement the self-organization algorithm in a distributed manner. In order to achieve extended sensor network lifetime, energy conservation was incorporated into the self-organization algorithm by incorporating a sleep-wakeup management mechanism with a novel cross layer adaptive wakeup probability adjustment scheme. The simulation results confirm that the developed self-organization algorithm provides satisfactory target tracking performance. Moreover, the energy saving analysis confirms the effectiveness of the cross layer power management scheme in achieving extended sensor network lifetime without degrading the target tracking performance.

Host and Network Optimizations for Performance Enhancement and Energy Efficiency in Data Center Networks

Modern data centers host hundreds of thousands of servers to achieve economies of scale. Such a huge number of servers create challenges for the data center network (DCN) to provide proportionally large bandwidth. In addition, the deployment of virtual machines (VMs) in data centers raises the requirements for efficient resource allocation and find-grained resource sharing. Further, the large number of servers and switches in the data center consume significant amounts of energy. Even though servers become more energy efficient with various energy saving techniques, DCN still accounts for 20% to 50% of the energy consumed by the entire data center. The objective of this dissertation is to enhance DCN performance as well as its energy efficiency by conducting optimizations on both host and network sides. First, as the DCN demands huge bisection bandwidth to interconnect all the servers, we propose a parallel packet switch (PPS) architecture that directly processes variable length packets without segmentation-and-reassembly (SAR). The proposed PPS achieves large bandwidth by combining switching capacities of multiple fabrics, and it further improves the switch throughput by avoiding padding bits in SAR. Second, since certain resource demands of the VM are bursty and demonstrate stochastic nature, to satisfy both deterministic and stochastic demands in VM placement, we propose the Max-Min Multidimensional Stochastic Bin Packing (M3SBP) algorithm. M3SBP calculates an equivalent deterministic value for the stochastic demands, and maximizes the minimum resource utilization ratio of each server. Third, to provide necessary traffic isolation for VMs that share the same physical network adapter, we propose the Flow-level Bandwidth Provisioning (FBP) algorithm. By reducing the flow scheduling problem to multiple stages of packet queuing problems, FBP guarantees the provisioned bandwidth and delay performance for each flow. Finally, while DCNs are typically provisioned with full bisection bandwidth, DCN traffic demonstrates fluctuating patterns, we propose a joint host-network optimization scheme to enhance the energy efficiency of DCNs during off-peak traffic hours. The proposed scheme utilizes a unified representation method that converts the VM placement problem to a routing problem and employs depth-first and best-fit search to find efficient paths for flows.

A molecular and stable isotopic approach to investigate the importance of algal and detrital energy pathways in a freshwater marsh

The relative importance of algal and detrital energy pathways remains a central question in wetlands ecology. We used bulk stable isotope analysis and fatty acid composition to investigate the relative contributions of periphyton (algae) and floc (detritus) in a freshwater wetland with the goal of determining the inputs of these resource pools to lower trophic-level consumers. All animal samples revealed fatty acid markers indicative of both microbial (detrital) and algal origins, though the relative contributions varied among species. Vascular plant markers were in low abundance in most consumers. Detritivory is important for chironomids and amphipods, as demonstrated by the enhanced bacterial fatty acids present in both consumers, while algal resources, in the form of periphyton, likely support ephemeropteran larvae. Invertebrates such as amphipods and grass shrimp appear to be important resources for small omnivorous fish, while Poecilia latipinna appear to strongly use periphyton and Ephemeroptera larvae as food sources. Both P. latipinna and Lepomis spp. assimilated small amounts of vascular plant debris, possibly due to unintentional ingestion of floc while foraging for invertebrates and insect larvae. Physid snails, Haitia spp., were characterized by considerably different fatty acid compositions than other taxa examined, and likely play a unique role in Everglades’ food webs.

An energy based nanomechnical properties evaluation method for cementitious materials

Advances in multiscale material modeling of structural concrete have created an upsurge of interest in the accurate evaluation of mechanical properties and volume fractions of its nano constituents. The task is accomplished by analyzing the response of a material to indentation, obtained as an outcome of a nanoindentation experiment, using a procedure called the Oliver and Pharr (OP) method. Despite its widespread use, the accuracy of this method is often questioned when it is applied to the data from heterogeneous materials or from the materials that show pile-up and sink-in during indentation, which necessitates the development of an alternative method. ^ In this study, a model is developed within the framework defined by contact mechanics to compute the nanomechanical properties of a material from its indentation response. Unlike the OP method, indentation energies are employed in the form of dimensionless constants to evaluate model parameters. Analysis of the load-displacement data pertaining to a wide range of materials revealed that the energy constants may be used to determine the indenter tip bluntness, hardness and initial unloading stiffness of the material. The proposed model has two main advantages: (1) it does not require the computation of the contact area, a source of error in the existing method; and (2) it incorporates the effect of peak indentation load, dwelling period and indenter tip bluntness on the measured mechanical properties explicitly. ^ Indentation tests are also carried out on samples from cement paste to validate the energy based model developed herein by determining the elastic modulus and hardness of different phases of the paste. As a consequence, it has been found that the model computes the mechanical properties in close agreement with that obtained by the OP method; a discrepancy, though insignificant, is observed more in the case of C-S-H than in the anhydrous phase. Nevertheless, the proposed method is computationally efficient, and thus it is highly suitable when the grid indentation technique is required to be performed. In addition, several empirical relations are developed that are found to be crucial in understanding the nanomechanical behavior of cementitious materials.^

Optimizing storage and memory systems for energy and performance

Electrical energy is an essential resource for the modern world. Unfortunately, its price has almost doubled in the last decade. Furthermore, energy production is also currently one of the primary sources of pollution. These concerns are becoming more important in data-centers. As more computational power is required to serve hundreds of millions of users, bigger data-centers are becoming necessary. This results in higher electrical energy consumption. Of all the energy used in data-centers, including power distribution units, lights, and cooling, computer hardware consumes as much as 80%. Consequently, there is opportunity to make data-centers more energy efficient by designing systems with lower energy footprint. Consuming less energy is critical not only in data-centers. It is also important in mobile devices where battery-based energy is a scarce resource. Reducing the energy consumption of these devices will allow them to last longer and re-charge less frequently. Saving energy in computer systems is a challenging problem. Improving a system's energy efficiency usually comes at the cost of compromises in other areas such as performance or reliability. In the case of secondary storage, for example, spinning-down the disks to save energy can incur high latencies if they are accessed while in this state. The challenge is to be able to increase the energy efficiency while keeping the system as reliable and responsive as before. This thesis tackles the problem of improving energy efficiency in existing systems while reducing the impact on performance. First, we propose a new technique to achieve fine grained energy proportionality in multi-disk systems; Second, we design and implement an energy-efficient cache system using flash memory that increases disk idleness to save energy; Finally, we identify and explore solutions for the page fetch-before-update problem in caching systems that can: (a) control better I/O traffic to secondary storage and (b) provide critical performance improvement for energy efficient systems.

High energy three -body breakup of three -nucleon systems

The main goal of this dissertation was to study two- and three-nucleon Short Range Correlations (SRCs) in high energy three-body breakup of 3He nucleus in 3He(e, e'NN) N reaction. SRCs are characterized by quantum fluctuations in nuclei during which constituent nucleons partially overlap with each other. ^ A theoretical framework is developed within the Generalized Eikonal Approximation (GEA) which upgrades existing medium-energy methods that are inapplicable for high momentum and energy transfer reactions. High momentum and energy transfer is required to provide sufficient resolution for probing SRCs. GEA is a covariant theory which is formulated through the effective Feynman diagrammatic rules. It allows self-consistent calculation of single and double re-scatterings amplitudes which are present in three-body breakup processes. The calculations were carried out in detail and the analytical result for the differential cross section of 3He(e, e'NN)N reaction was derived in a form applicable for programming and numerical calculations. The corresponding computer code has been developed and the results of computation were compared to the published experimental data, showing satisfactory agreement for a wide range of values of missing momenta. ^ In addition to the high energy approximation this study exploited the exclusive nature of the process under investigation to gain more information about the SRCs. The description of the exclusive 3He( e, e'NN)N reaction has been done using the formalism of the nuclear decay function, which is a practically unexplored quantity and is related to the conventional spectral function through the integration of the phase space of the recoil nucleons. Detailed investigation showed that the decay function clearly exhibits the main features of two- and three-nucleon correlations. Four highly practical types of SRCs in 3He nucleus were discussed in great detail for different orders of the final state re-interactions using the decay function as an unique identifying tool. ^ The overall conclusion in this dissertation suggests that the investigation of the decay function opens up a completely new venue in studies of short range nuclear properties. ^

Deployment of a Hybrid Multicast Switch in Energy-Aware Data Center Network: A Case of Fat-Tree Topology

Recently, energy efficiency or green IT has become a hot issue for many IT infrastructures as they attempt to utilize energy-efficient strategies in their enterprise IT systems in order to minimize operational costs. Networking devices are shared resources connecting important IT infrastructures, especially in a data center network they are always operated 24/7 which consume a huge amount of energy, and it has been obviously shown that this energy consumption is largely independent of the traffic through the devices. As a result, power consumption in networking devices is becoming more and more a critical problem, which is of interest for both research community and general public. Multicast benefits group communications in saving link bandwidth and improving application throughput, both of which are important for green data center. In this paper, we study the deployment strategy of multicast switches in hybrid mode in energy-aware data center network: a case of famous fat-tree topology. The objective is to find the best location to deploy multicast switch not only to achieve optimal bandwidth utilization but also to minimize power consumption. We show that it is possible to easily achieve nearly 50% of energy consumption after applying our proposed algorithm.

Weak electron scattering for the 3He-3H transition and the weak nuclear form factors

We calculate the differential cross section for weak electron scattering reaction, e + 3He-' 3H + ve, for energies from 100 MeV to 6 GeV as a function of outgoing nucleus angle from 0 to n/2 radians. We find that the differential cross section at low [q2] increases with electron energy from 0.1 GeV to 6.0 GeV, such that the peak value at 6.0 GeV is approximately 3.2 x 10-40 cm 2 / ster, a factor of 10 larger than the peak value at 0.1 GeV. We also find that the width of the peak falls very rapidly with increasing electron energy. At high [q2] we find that the differential cross section falls by approximately three orders of magnitude making experimental observation at this time unlikely. The contributions of the individual form factors are obtained for electron energies of 0.5GeV and 2.0 GeV. It is found that at low [q2] the form factors, FA(q2) and Fv(q2), make contributions of similar size to the differential cross section and might be simultaneously determined , but for the case of FM(q2) we find that the contribution is too small to determine. It is also found that at large [q2] values, the contribution of FM(q2) is substantially enhanced , but that the cross section is probably too small to enable a direct determination of FM(q2).

High energy three-body breakup of three-nucleon systems

The main goal of this dissertation was to study two- and three-nucleon Short Range Correlations (SRCs) in high energy three-body breakup of 3He nucleus in 3He(e, e'NN)N reaction. SRCs are characterized by quantum fluctuations in nuclei during which constituent nucleons partially overlap with each other. A theoretical framework is developed within the Generalized Eikonal Approximation (GEA) which upgrades existing medium-energy methods that are inapplicable for high momentum and energy transfer reactions. High momentum and energy transfer is required to provide sufficient resolution for probing SRCs. GEA is a covariant theory which is formulated through the effective Feynman diagrammatic rules. It allows self-consistent calculation of single and double re-scatterings amplitudes which are present in three-body breakup processes. The calculations were carried out in detail and the analytical result for the differential cross section of 3He(e, e'NN)Nreaction was derived in a form applicable for programming and numerical calculations. The corresponding computer code has been developed and the results of computation were compared to the published experimental data, showing satisfactory agreement for a wide range of values of missing momenta. In addition to the high energy approximation this study exploited the exclusive nature of the process under investigation to gain more information about the SRCs. The description of the exclusive 3He(e, e'NN)N reaction has been done using the formalism of the nuclear decay function, which is a practically unexplored quantity and is related to the conventional spectral function through the integration of the phase space of the recoil nucleons. Detailed investigation showed that the decay function clearly exhibits the main features of two- and three-nucleon correlations. Four highly practical types of SRCs in 3He nucleus were discussed in great detail for different orders of the final state re-interactions using the decay function as an unique identifying tool. The overall conclusion in this dissertation suggests that the investigation of the decay function opens up a completely new venue in studies of short range nuclear properties.