NeCTAR Research Cloud Research Articles

Classification of Twitter users with eating disorder engagement: Learning from the biographies

Individuals with an Eating Disorder (ED) are typically reluctant to seek help via traditional means (e.g., psychologists). However, recent evidence suggests that many individuals seek assistance via social media for weight and diet related concerns. Sophisticated approaches are needed to better distinguish those who may be in need of help for an ED from those who are simply commenting on ED in online social environments. In order to facilitate effective communication between individuals with or at-risk of an ED and healthcare professionals, this research exploits a deep learning model to differentiate the users with ED engagement (e.g., ED sufferers, healthcare professionals or communicators) over social media. For this purpose, a collection of Twitter data is compiled using Twitter application programming interface (API) on the Australian Research Data Commons (ARDC) Nectar research cloud. After collecting 1,400,000 Twitter biographies in total, a subset of 4000 biographies are annotated manually. This annotation enables the differentiation of users engaged with ED-focused language on social media into five categories: ED-user, healthcare professional, communicator, healthcare professional-communicator, and other. Based on these annotated categories, a predictive deep learning model based on bidirectional encoder representations from transformers (BERT) and long short-term memory (LSTM) is developed. The model achieves an F1 score of 98.19% and an accuracy of 98.37%. It demonstrates the viability of detecting the individuals with possible ED risk and distinguishes them from other categories using their biography data. We further conducted a network analysis for investigating the communication network between these categories. Our analysis shows that ED-users are more secretive and self-protective, whereas the healthcare professionals and communicators frequently interact with each other and a wide range of other people. To the best of our knowledge, our research is the first of its kind for identifying the different user categories engaged with ED-focused communications on social media.

Performance evaluation of Hyperledger Fabric-enabled framework for pervasive peer-to-peer energy trading in smart Cyber–Physical Systems

The in-depth collaboration of Cyber–Physical Systems (CPSs) and smart grids constitute the novel paradigm of distributed energy trading, in which computation and process control are managed in an adaptive Peer-to-Peer (P2P) manner. To further strengthen this collaboration, Hyperledger Fabric (HF) can be prominently considered as a mean to implement next-generation secure and intelligent communication. However, implementing real-world applications on this platform may concern performance issues. For the constructive exploration of these issues, initially, we design a novel P2P energy trading framework for improving resource utilization and consequently addressing the impending electricity crisis challenge. Thenceforward, we evaluate the results based on the different system operational parameters for establishing a proof-of-concept. For determining performance bottlenecks and best-configuration, these results are investigated independently by using the Nectar Research Cloud, thereby sustaining scalability. The proposed evaluation approach will largely contribute to determining the system operational-level parameters of enterprise applications that will utilize the HF platform as their communication tool-support. In addition, a benchmark is presented based on the Hyperledger Caliper tool to facilitate application designers and developers in the form of selecting an appropriate implementation model across the two latest stable HF model versions. The illustrative CPS-enabled energy trading scenario corroborates the feasibility of the proposed framework to foster the development of HF-assisted smart P2P energy trading mechanisms.

SRA Down Under: Cache and Analysis Platform for Infectious Disease.

SRA, NCBI's Sequence Read Archive, is a valuable resource holding a near definitive collection of the world's collective sequenced reads for academic purposes. Increasingly, these reads are being used for both basic research and clinical investigations. When time is a critical factor in analysis, such as during bacterial outbreaks, the geographical separation between Australia and the offshore NCBI SRA servers can result in significant delays that may have adverse clinical outcomes. To address this, Queensland Genomics commissioned a pilot program for the establishment of a local Australian SRA Cache. Utilizing the hosting capabilities of the NeCTAR Research Cloud, QRIScloud's HTC infrastructure and the MeDiCI data fabric as a storage solution, and the software stack of Cromwell for workflow management, PostgreSQL database for sample and job metadata, and a coordinator Python Flask application, a local cache of seventeen bacterial species was established. Furthermore, the workflow capabilities of Cromwell were leveraged to provide analysis solutions for cached sample data, including quality control and taxonomic profiling, and individual and multiple sample analysis. Moving forward to a broader rollout of increased bacterial species, it was found that the initial storage estimation did not keep up with the exponential increase sequencing reads uploaded to NCBI SRA, which while highlighting the increasing availability and importance in modern research, will need to be addressed.

Evaluating virtual hosted desktops for graphics-intensive astronomy

Visualisation of data is critical to understanding astronomical phenomena. Today, many instruments produce datasets that are too big to be downloaded to a local computer, yet many of the visualisation tools used by astronomers are deployed only on desktop computers. Cloud computing is increasingly used to provide a computation and simulation platform in astronomy, but it also offers great potential as a visualisation platform. Virtual hosted desktops, with graphics processing unit (GPU) acceleration, allow interactive, graphics-intensive desktop applications to operate co-located with astronomy datasets stored in remote data centres. By combining benchmarking and user experience testing, with a cohort of 20 astronomers, we investigate the viability of replacing physical desktop computers with virtual hosted desktops. In our work, we compare two Apple MacBook computers (one old and one new, representing hardware and opposite ends of the useful lifetime) with two virtual hosted desktops: one commercial (Amazon Web Services) and one in a private research cloud (the Australian Nectar Research Cloud). For two-dimensional image-based tasks and graphics-intensive three-dimensional operations -- typical of astronomy visualisation workflows -- we found that benchmarks do not necessarily provide the best indication of performance. When compared to typical laptop computers, virtual hosted desktops can provide a better user experience, even with lower performing graphics cards. We also found that virtual hosted desktops are equally simple to use, provide greater flexibility in choice of configuration, and may actually be a more cost-effective option for typical usage profiles.

Geo-aware erasure coding for high-performance erasure-coded storage clusters

Erasure code-based distributed storage systems are increasingly being used by storage providers for big data storage since they offer the same reliability as replication with a significant decrease in the amount of storage required. But, when it comes to a storage system with data nodes spread across a very large geographical area, the node’s recovery performance is affected by various factors that are both network and computation related. In this paper, we present a XOR-based code supplemented with the ideas of parity duplication and rack awareness that could be adopted in such storage clusters to improve the recovery performance during node failures and compare it with popular implementations of erasure codes, namely Facebook’s Reed-Solomon codes and XORBAS local recovery codes. The code performance along with the proposed ideas are evaluated on a geo-diverse cluster deployed on the NeCTAR research cloud. We also present a scheme for intelligently placing blocks of coded storage depending on the design of the code, inspired by local reconstruction codes. The sum of all these propositions could offer a better solution for applications that are deployed on coded storage systems that are geographically distributed, in which storage constraints make triple replication not affordable, at the same time ensuring minimal recovery time is a strict requirement.

Migration of a SCADA system to IaaS clouds \u2013 a case study

SCADA systems allow users to monitor and/or control physical devices, processes, and events remotely and in real-time. As these systems are critical to industrial processes, they are often run on highly reliable and dedicated hardware. Moving these SCADA systems to an Infrastructure as a Service (IaaS) cloud allows for: cheaper deployments, system redundancy support, and increased uptime. The goal of this work was to present the results of our experimental study of moving/migrating a selected SCADA system to a cloud environment and present major lessons learned. To this end, EclipseSCADA was deployed to the NeCTAR research cloud using the “lift and shift” approach. Performance metrics of a unique nature and large scale of experimentation were collected from the deployed EclipseSCADA system under different loads to examine the effects cloud resources and public networks have on SCADA behavior.

Next-generation sequencing: a challenge to meet the increasing demand for training workshops in Australia

The widespread adoption of high-throughput next-generation sequencing (NGS) technology among the Australian life science research community is highlighting an urgent need to up-skill biologists in tools required for handling and analysing their NGS data. There is currently a shortage of cutting-edge bioinformatics training courses in Australia as a consequence of a scarcity of skilled trainers with time and funding to develop and deliver training courses. To address this, a consortium of Australian research organizations, including Bioplatforms Australia, the Commonwealth Scientific and Industrial Research Organisation and the Australian Bioinformatics Network, have been collaborating with EMBL-EBI training team. A group of Australian bioinformaticians attended the train-the-trainer workshop to improve training skills in developing and delivering bioinformatics workshop curriculum. A 2-day NGS workshop was jointly developed to provide hands-on knowledge and understanding of typical NGS data analysis workflows. The road show–style workshop was successfully delivered at five geographically distant venues in Australia using the newly established Australian NeCTAR Research Cloud. We highlight the challenges we had to overcome at different stages from design to delivery, including the establishment of an Australian bioinformatics training network and the computing infrastructure and resource development. A virtual machine image, workshop materials and scripts for configuring a machine with workshop contents have all been made available under a Creative Commons Attribution 3.0 Unported License. This means participants continue to have convenient access to an environment they had become familiar and bioinformatics trainers are able to access and reuse these resources.