AdFT: An Adaptive Framework for Fault Tolerance on Large Scale Systems using Application Malleability

Exascale systems of the future are predicted to have mean time between failures (MTBF) of less than one hour. Malleable applications, where the number of processors on which the applications execute can be changed during executions, can make use of their malleability to better tolerate high failure rates. We present AdFT, an adaptive fault tolerance framework for long running malleable applications to maximize application performance in the presence of failures. AdFT framework includes cost models for evaluating the benefits of various fault tolerance actions including checkpointing, live-migration and rescheduling, and runtime decisions for dynamically selecting the fault tolerance actions at different points of application execution to maximize performance. Simulations with real and synthetic failure traces show that our approach outperforms existing fault tolerance mechanisms for malleable applications yielding up to 23% improvement in application performance, and is effective even for petascale systems and beyond.

Nonuniform random geometric graphs with location-dependent radii

We propose a distribution-free approach to the study of random geometric graphs. The distribution of vertices follows a Poisson point process with intensity function n f(center dot), where n is an element of N, and f is a probability density function on R-d. A vertex located at x connects via directed edges to other vertices that are within a cut-off distance r(n)(x). We prove strong law results for (i) the critical cut-off function so that almost surely, the graph does not contain any node with out-degree zero for sufficiently large n and (ii) the maximum and minimum vertex degrees. We also provide a characterization of the cut-off function for which the number of nodes with out-degree zero converges in distribution to a Poisson random variable. We illustrate this result for a class of densities with compact support that have at most polynomial rates of decay to zero. Finally, we state a sufficient condition for an enhanced version of the above graph to be almost surely connected eventually.

Wave propagation in a porous composite beam: Porosity determination, location and quantification

A wave-based method is developed to quantify the defect due to porosity and also to locate the porous regions, in a composite beam-type structure. Wave propagation problem for a porous laminated composite beam is modeled using spectral finite element method (SFEM), based on the modified rule of mixture approach, which is used to include the effect of porosity on the stiffness and density of the composite beam structure. The material properties are obtained from the modified rule of mixture model, which are used in a conventional SFEM to develop a new model for solving wave propagation problems in porous laminated composite beam. The influence of the porosity content on the group speed and also the effect of variation in theses parameters on the time responses are studied first, in the forward problem. The change in the time responses with the change in the porosity of the structure is used as a parameter to find the porosity content in a composite beam. The actual measured response from a structure and the numerically obtained time responses are used for the estimation of porosity, by solving a nonlinear optimization problem. The effect of the length of the porous region (in the propagation direction), on the time responses, is studied. The damage force indicator technique is used to locate the porous region in a beam and also to find its length, using the measured wave propagation responses. (C) 2012 Elsevier Ltd. All rights reserved.

Effect of fault resistance and grid short circuit MVA on impedance seen by distance relays on lines fed from wind turbine generating units (WTGU)

This paper deals with line protection challenges experienced in system having substantial wind generation penetration. Two types of WTGU: Doubly Fed (DFIG) and Squirrel Cage (SCIG) Induction Generators are simulated and connected to grid with single circuit transmission line. The paper summarizes analytical investigations carried out on the impedance seen by distance relays by varying fault resistances and grid short circuit MVA, for the protection of such transmission lines during faults. The results are also compared with systems having conventional synchronous machine connected to the grid.

A GIS (geographical information system)-based spatial data mining approach for optimal location and capacity planning of distributed biomass power generation facilities: A case study of Tumkur district, India

This paper primarily intends to develop a GIS (geographical information system)-based data mining approach for optimally selecting the locations and determining installed capacities for setting up distributed biomass power generation systems in the context of decentralized energy planning for rural regions. The optimal locations within a cluster of villages are obtained by matching the installed capacity needed with the demand for power, minimizing the cost of transportation of biomass from dispersed sources to power generation system, and cost of distribution of electricity from the power generation system to demand centers or villages. The methodology was validated by using it for developing an optimal plan for implementing distributed biomass-based power systems for meeting the rural electricity needs of Tumkur district in India consisting of 2700 villages. The approach uses a k-medoid clustering algorithm to divide the total region into clusters of villages and locate biomass power generation systems at the medoids. The optimal value of k is determined iteratively by running the algorithm for the entire search space for different values of k along with demand-supply matching constraints. The optimal value of the k is chosen such that it minimizes the total cost of system installation, costs of transportation of biomass, and transmission and distribution. A smaller region, consisting of 293 villages was selected to study the sensitivity of the results to varying demand and supply parameters. The results of clustering are represented on a GIS map for the region.

EFFECT OF OFF-STOICHIOMETRY AND TERNARY ADDITIONS ON PLANAR FAULT ENERGIES IN Ni3Al

First principles calculations were done to evaluate the lattice parameter, cohesive energy and stacking fault energies of ordered gamma' (Ll(2)) precipitates in superalloys as a function of composition. It was found that addition of Ti and Ta lead to an increase in lattice parameter and decrease in cohesive energy, while Ni antisites had the opposite effect. Ta and Ti addition to stoichiometric Ni3Al resulted in an initial increase in the energies of APB((111)), CSF(111), APB((001)) and SISF(111). However, at higher concentrations, the fault energies decreased. Addition of Ni antisites decreased the energy of all four faults monotonically. A model based on nearest neighbor bonding was used for Ni-3(Al, Ta), Ni-3(Al, Ti) and Ni-3(Al, Ni) pseudo-binary systems and extended to pseudo- ternary Ni-3(Al, Ta, Ni) and Ni-3(Al, Ti, Ni) systems. Recipes were developed for predicting lattice parameters, cohesive energies and fault energies in pseudo- ternary systems on the basis of coefficients derived from simpler pseudobinary systems. The model predictions were found to be in good agreement with first principles calculations for lattice parameters, cohesive energies, and energies of APB((111)) and CSF(111).

Fault-tolerant secret key generation

Mobile nodes observing correlated data communicate using an insecure bidirectional switch to generate a secret key, which must remain concealed from the switch. We are interested in fault-tolerant secret key rates, i.e., the rates of secret key generated even if a subset of nodes drop out before the completion of the communication protocol. We formulate a new notion of fault-tolerant secret key capacity, and present an upper bound on it. This upper bound is shown to be tight when the random variables corresponding to the observations of nodes are exchangeable. Further, it is shown that one round of interaction achieves the fault-tolerant secret key capacity in this case. The upper bound is also tight for the case of a pairwise independent network model consisting of a complete graph, and can be attained by a noninteractive protocol.

A Reduced Device-Count Hybrid Multilevel Inverter Topology with single DC Source and Improved Fault Tolerance

A new hybrid multilevel power converter topology is presented in this paper. The proposed power converter topology uses only one DC source and floating capacitors charged to asymmetrical voltage levels, are used for generating different voltage levels. The SVPWM based control strategy used in this converter maintains the capacitor voltages at the required levels in the entire modulation range including the over-modulation region. For the voltage levels: nine and above, the number of components required in the proposed topology is significantly lower, compared to the conventional multilevel inverter topologies. The number of capacitors required in this topology reduces drastically compared to the conventional flying capacitor topology, when the number of levels in the inverter output increases. This topology has better fault tolerance, as it is capable of operating with reduced number of levels, in the entire modulation range, in the event of any failure in the H-bridges. The transient as well as the steady state performance of the nine-level version of the proposed topology is experimentally verified in the entire modulation range including the over-modulation region.

Active Dendrites Regulate Spectral Selectivity in Location-Dependent Spike Initiation Dynamics of Hippocampal Model Neurons

How does the presence of plastic active dendrites in a pyramidal neuron alter its spike initiation dynamics? To answer this question, we measured the spike-triggered average (STA) from experimentally constrained, conductance-based hippocampal neuronal models of various morphological complexities. We transformed the STA computed from these models to the spectral and the spectrotemporal domains and found that the spike initiation dynamics exhibited temporally localized selectivity to a characteristic frequency. In the presence of the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, the STA characteristic frequency strongly correlated with the subthreshold resonance frequency in the theta frequency range. Increases in HCN channel density or in input variance increased the STA characteristic frequency and its selectivity strength. In the absence of HCN channels, the STA exhibited weak delta frequency selectivity and the characteristic frequency was related to the repolarization dynamics of the action potentials and the recovery kinetics of sodium channels from inactivation. Comparison of STA obtained with inputs at various dendritic locations revealed that nonspiking and spiking dendrites increased and reduced the spectrotemporal integration window of the STA with increasing distance from the soma as direct consequences of passive filtering and dendritic spike initiation, respectively. Finally, the presence of HCN channels set the STA characteristic frequency in the theta range across the somatodendritic arbor and specific STA measurements were strongly related to equivalent transfer-impedance-related measurements. Our results identify explicit roles for plastic active dendrites in neural coding and strongly recommend a dynamically reconfigurable multi-STA model to characterize location-dependent input feature selectivity in pyramidal neurons.

Deformation of nanograined Ni-60Co alloy with low stacking fault energy

The evolution of deformation texture in a Ni-60Co alloy with low stacking fault energy and a grain size in the nanometre range has been investigated. The analyses of texture and microstructure suggest different mechanisms of deformation in nanocrystalline as compared to microcrystalline Ni-60Co alloy. In nanocrystalline material, the mechanism responsible for texture formation has been identified as partial slip, whereas in microcrystalline material, a characteristic texture forms due to twinning and shear banding.

Fault Attacks on Pairing-Based Protocols Revisited

Several papers have studied fault attacks on computing a pairing value e(P, Q), where P is a public point and Q is a secret point. In this paper, we observe that these attacks are in fact effective only on a small number of pairing-based protocols, and that too only when the protocols are implemented with specific symmetric pairings. We demonstrate the effectiveness of the fault attacks on a public-key encryption scheme, an identity-based encryption scheme, and an oblivious transfer protocol when implemented with a symmetric pairing derived from a supersingular elliptic curve with embedding degree 2.

Role of stacking fault energy on texture evolution revisited

Three materials, pure aluminium, Al-4 wt.% Mg, alpha-brass have been chosen to understand the evolution of texture and microstructure during rolling. Pure Al develops a strong copper-type rolling texture and the deformation is entirely slip dominated. In Al-4Mg alloy, texture is copper-type throughout the deformation. The advent of Cu-type shear bands in the later stages of deformation has a negligible effect on the final texture. alpha-brass shows a characteristic brass-type texture from the early stages of rolling. Extensive twinning in the intermediate stages of deformation (epsilon(t) similar to 0.5) causes significant texture reorientation towards alpha-fiber. Beyond 40% reduction, deformation is dominated by Bs-type shear bands, and the banding coincides with the evolution of <111>parallel to ND components. The crystallites within the bands preferentially show <110>parallel to ND components. The absence of the Cu component throughout the deformation process indicates that, for the evolution of brass-type texture, the presence of Cu component is not a necessary condition. The final rolling texture is a synergistic effect of deformation twinning and shear banding.

Engineering nanostructured polymer blends with controlled nanoparticle location for excellent microwave absorption: a compartmentalized approach

In order to obtain better materials, control over the precise location of nanoparticles is indispensable. It is shown here that ordered arrangements of nanoparticles, possessing different characteristics (electrical/ magnetic dipoles), in the blend structure can result in excellent microwave absorption. This is manifested from a high reflection loss of ca. -67 dB for the best blend structure designed here. To attenuate electromagnetic radiation, the key parameters of high electrical conductivity and large dielectric/magnetic loss are targeted here by including a conductive material multiwall carbon nanotubes, MWNTs], ferroelectric nanostructured material with associated relaxations in the GHz frequency barium titanate, BT] and lossy ferromagnetic nanoparticles nickel ferrite, NF]. In this study, bi-continuous structures were designed using 50/50 (by wt) blends of polycarbonate (PC) and polyvinylidene fluoride (PVDF). The MWNTs were modified using an electron acceptor molecule, a derivative of perylenediimide, which facilitates p-p stacking with the nanotubes and stimulates efficient charge transport in the blends. The nanoscopic materials have specific affinity towards the PVDF phase. Hence, by introducing surface-active groups, an ordered arrangement can be tailored. To accomplish this, both BT and NF were first hydroxylated followed by the introduction of amine-terminal groups on the surface. The latter facilitated nucleophilic substitution reactions with PC and resulted in their precise location. In this study, we have shown for the first time that by a compartmentalized approach, superior EM attenuation can be achieved. For instance, when the nanoparticles were localized exclusively in the PVDF phase or in both the phases, the minimum reflection losses were ca. -18 dB (for the MWNT/BT mixture) and -29 dB (for the MWNT/NF mixture), and the shielding occurred primarily through reflection. Interestingly, by adopting the compartmentalized approach wherein the lossy materials were in the PC phase and the conductive materials (MWNT) were in the PVDF phase, outstanding reflection losses of ca. -57 dB (for the BT and MWNT combination) and -67 dB (for the NF and MWNT combination) were noted and the shielding occurred primarily through absorption. Thus, the approach demonstrates that nanoscopic structuring in the blends can be achieved under macroscopic processing conditions and this strategy can further be explored to design microwave absorbers.

Fault Tolerance on Large Scale Systems using Adaptive Process Replication

Exascale systems of the future are predicted to have mean time between failures (MTBF) of less than one hour. At such low MTBFs, employing periodic checkpointing alone will result in low efficiency because of the high number of application failures resulting in large amount of lost work due to rollbacks. In such scenarios, it is highly necessary to have proactive fault tolerance mechanisms that can help avoid significant number of failures. In this work, we have developed a mechanism for proactive fault tolerance using partial replication of a set of application processes. Our fault tolerance framework adaptively changes the set of replicated processes periodically based on failure predictions to avoid failures. We have developed an MPI prototype implementation, PAREP-MPI that allows changing the replica set. We have shown that our strategy involving adaptive process replication significantly outperforms existing mechanisms providing up to 20 percent improvement in application efficiency even for exascale systems.