Heavy-tailed limits for medium size jobs and comparison scheduling

We study the conditional sojourn time distributions of processor sharing (PS), foreground background processor sharing (FBPS) and shortest remaining processing time first (SRPT) scheduling disciplines on an event where the job size of a customer arriving in stationarity is smaller than exactly k≥0 out of the preceding m≥k arrivals. Then, conditioning on the preceding event, the sojourn time distribution of this newly arriving customer behaves asymptotically the same as if the customer were served in isolation with a server of rate (1−ρ)/(k+1) for PS/FBPS, and (1−ρ) for SRPT, respectively, where ρ is the traffic intensity. Hence, the introduced notion of conditional limits allows us to distinguish the asymptotic performance of the studied schedulers by showing that SRPT exhibits considerably better asymptotic behavior for relatively smaller jobs than PS/FBPS. Inspired by the preceding results, we propose an approximation to the SRPT discipline based on a novel adaptive job grouping mechanism that uses relative size comparison of a newly arriving job to the preceding m arrivals. Specifically, if the newly arriving job is smaller than k and larger than m−k of the previous m jobs, it is routed into class k. Then, the classes of smaller jobs are served with higher priorities using the static priority scheduling. The good performance of this mechanism, even for a small number of classes m+1, is demonstrated using the asymptotic queueing analysis under the heavy-tailed job requirements. We also discuss refinements of the comparison grouping mechanism that improve the accuracy of job classification at the expense of a small additional complexity. This work is supported by NSF Grant 0615126.     

Insights into the scalability of magnetostrictive ultrasound technology for water treatment applications

To date, the successful application of large scale ultrasound in water treatment has been a challenge. Magnetostrictive ultrasound technologies for constructing a large-scale water treatment system are proposed in this study. Comprehensive energy evaluation of the proposed system was conducted. The effects of chosen waveform, scalability and reactor design on the performance of the system were explored using chemical dosimetry. Of the fundamental waveforms tested; sine, triangle and square, the highest chemical yield resulted from the square wave source. Scaling up from the 0.5 L bench-scale system to the 15 L large-scale unit resulted in a gain of approximately 50% in sonochemical efficiency (SE) for the system. The use of a reactor tank with 45° inclined sides further increased SE of the system by 70%. The ability of the large scale system in removing contaminants from natural water samples was also investigated. The results revealed that the large-scale unit was capable of achieving a maximum removal of microbes and dissolved organic carbon (DOC) of 35% and 5.7% respectively at a power density approximately 3.9 W/L. The results of this study suggest that magnetostrictive ultrasound technology excited with square wave has the potential to be competitive in the water treatment industry.     

A parallel multi-configuration self-consistent field algorithm

A scalable multi-configuration self-consistent field (MCSCF) algorithm is described. The method for optimizing the orbital and configurational parameters is based upon the two-step Newton–Raphson approach with an augmented orbital Hessian matrix. A single copy of the two-electron integrals in the molecular orbital basis is distributed over the memory of all processors. Storage of the augmented Hessian is avoided by re-computing its elements as needed. A replicated data approach is used to parallelize the configuration interaction step. Scalability to 1024 processors is demonstrated.     

A unified trust management strategy for content sharing in Peer-to-Peer networks

Content sharing has emerged as a popular application in Peer-to-Peer (P2P) networks. The increasing number of free-riders and untrustworthy peers in content sharing network reduces the credibility of the shared contents. Also, the malicious peers pose the danger of security violation and bandwidth wastage by sharing polluted or infected contents in the network for their personal benefits. In this paper, we propose a unified trust management strategy by combining trust and security models with reputation based preferences for efficient content sharing in P2P networks. The problems in P2P content sharing, like free-riding, content pollution, content poisoning, peer availability, content availability, DoS attack, and flash crowd are considered in our unified trust management model. The model provides controlled scalability of the overlay network, and concurrently ensures the availability of contents for longer time in the network. The provider initiating content sharing is protected from Denial-of-Service (DoS) attack through restricted network size and interspersing the provider within a set of distributed peers. Performance measures, such as scalability, content availability, flash crowd and prevention of polluted or infected content distribution are compared with BitTorrent in heterogeneous environment. The advantage of this model is that it can be deployed easily in BitTorrent-like P2P networks. Simulation results show that controlled scalability helps in efficient utilization of bandwidth and protection for the providers, ultimately ensuring trust on the P2P network and the content shared.     

Divide and Recombine Approaches for Fitting Smoothing Spline Models with Large Datasets

Spline smoothing is a widely used nonparametric method that allows data to speak for themselves. Due to its complexity and flexibility, fitting smoothing spline models is usually computationally intensive which may become prohibitive with large datasets. To overcome memory and CPU limitations, we propose four divide and recombine (D&R) approaches for fitting cubic splines with large datasets. We consider two approaches to divide the data: random and sequential. For each approach of division, we consider two approaches to recombine. These D&R approaches are implemented in parallel without communication. Extensive simulations show that these D&R approaches are scalable and have comparable performance as the method that uses the whole data. The sequential D&R approaches are spatially adaptive which lead to better performance than the method that uses the whole data when the underlying function is spatially inhomogeneous.     

Blockchain based Scalable model for Secure Dynamic Spectrum Access

Conventional Cooperative spectrum sensing techniques either suffer from single point of failure attack or lack in providing incentives to users which makes them incompatible for Wireless Service Provider (WSP). We propose a dynamic spectrum access framework for WSP which gives prominence to automated sensing and sharing with the use of blockchain. In this system, the opportunity of spectrum access is first examined by sensor nodes and the access right is then allocated to the users when their transactions to WSP are authenticated in a decentralized manner. Apart from using blockchain as a reliable platform for automatic enforcement of spectrum sensing, we propose a novel mechanism for securing our network from the threats designed primarily for Cognitive Radio Networks. In addition to this, our proposed approach enhances the scalability of blockchain networks by using the sidechains for storing data and checkpointing it onto main chain after periodic intervals of time. Extensive simulations in Octave indicate superior performance offered by our proposed model.     

Toward Scalable Nanofluidic Osmotic Power Generation from Hypersaline Water Sources with a Metal–Organic Framework Membrane

Nanofluidic membranes have shown great promise in harvesting osmotic energy but its scalablity remains challenging since most studies only tested with a membrane area of ≈10⁻² mm² or smaller. We demonstrate that metal-organic-framework membranes with subnanometer pores can be used for scalable osmotic power generation from hypersaline water sources. Our membrane can be scaled up to a few mm², and the power density can be stabilized at 1.7 W m⁻². We reveal that the key is to improve the out-of-membrane conductance while keeping the membrane's charge selectivity, contradicting the previous conception that the ionic conductivity of the membrane plays the dominating role. We highlight that subnanometer pores are essential to ensure the charge selectivity in hypersaline water sources. Our results suggest the importance to engineer the interplay between the in-membrane and out-of-membrane ion transport properties for scalable osmotic power generation.