Artificial intelligence (AI) server infrastructure has been built to support AI applications and handle data-intensive workloads. AI server infrastructure is the essential building blocks, and errors ...in AI server infrastructure may lead to fatal consequences to any AI applications built upon it. Compared to traditional software, software for AI server infrastructure is more configurable, and thus more likely to have configuration errors that might prevent correct software behaviors. Previous work on misconfiguration diagnosis requires sufficient execution history or manual intervention, and can hardly diagnose potential misconfigurations which are not triggered at launching. In this paper, we propose a real-time method to address these issues. Specifically, we combine program analysis and real-time log parsing to diagnose configuration errors. It maps each configuration option to the log code by applying program slicing only once, and parses real-time logs during the operation of the AI server without manual intervention. We evaluate the effectiveness of our approach on the core components of Hadoop, an exemplar AI Server Infrastructure Software. The results show that our method mapped more than 80% of the configuration options to log outputs, identified 90% of the configuration read sites as the slicing seeds, and successfully diagnosed about 10% configuration errors that can not be addressed by previous studies.
Abstract
Arm selection, the preferential expression of a 3′ or 5′ mature microRNA (miRNA), is a highly dynamic and tissue-specific process. Time-dependent expression shifts or switches between the ...arms are also relevant for human diseases. We present miRSwitch, a web server to facilitate the analysis and interpretation of arm selection events. Our species-independent tool evaluates pre-processed small non-coding RNA sequencing (sncRNA-seq) data, i.e. expression matrices or output files from miRNA quantification tools (miRDeep2, miRMaster, sRNAbench). miRSwitch highlights potential changes in the distribution of mature miRNAs from the same precursor. Group comparisons from one or several user-provided annotations (e.g. disease states) are possible. Results can be dynamically adjusted by choosing from a continuous range of highly specific to very sensitive parameters. Users can compare potential arm shifts in the provided data to a human reference map of pre-computed arm shift frequencies. We created this map from 46 tissues and 30 521 samples. As case studies we present novel arm shift information in a Alzheimer’s disease biomarker data set and from a comparison of tissues in Homo sapiens and Mus musculus. In summary, miRSwitch offers a broad range of customized arm switch analyses along with comprehensive visualizations, and is freely available at: https://www.ccb.uni-saarland.de/mirswitch/.
Abstract
Next-generation sequencing has paved the way for the reconstruction of genome-scale metabolic networks as a powerful tool for understanding metabolic circuits in any organism. However, the ...visualization and extraction of knowledge from these large networks comprising thousands of reactions and metabolites is a current challenge in need of user-friendly tools. Here we present Fluxer (https://fluxer.umbc.edu), a free and open-access novel web application for the computation and visualization of genome-scale metabolic flux networks. Any genome-scale model based on the Systems Biology Markup Language can be uploaded to the tool, which automatically performs Flux Balance Analysis and computes different flux graphs for visualization and analysis. The major metabolic pathways for biomass growth or for biosynthesis of any metabolite can be interactively knocked-out, analyzed and visualized as a spanning tree, dendrogram or complete graph using different layouts. In addition, Fluxer can compute and visualize the k-shortest metabolic paths between any two metabolites or reactions to identify the main metabolic routes between two compounds of interest. The web application includes >80 whole-genome metabolic reconstructions of diverse organisms from bacteria to human, readily available for exploration. Fluxer enables the efficient analysis and visualization of genome-scale metabolic models toward the discovery of key metabolic pathways.
BCube is one kind of important data center networks. Hamiltonicity and Hamiltonian connectivity have significant applications in communication networks. So far, there have been many results ...concerning fault-tolerant Hamiltonicity and fault-tolerant Hamiltonian connectivity in some data center networks. However, these results only consider faulty edges and faulty servers. In this paper, we study the fault-tolerant Hamiltonicity and the fault-tolerant Hamiltonian connectivity of
BCube
(
n, k
) under considering faulty servers, faulty links/edges, and faulty switches. For any integers
n
≥ 2 and
k
≥ 0, let
BC
n,k
be the logic structure of
BCube
(
n, k
) and
F
be the union of faulty elements of
BC
n,k
. Let
f
v
,
f
e
, and
f
s
be the number of faulty servers, faulty edges, and faulty switches of
BCube
(
n, k
), respectively. We show that
BC
n,k
−
F
is fault-tolerant Hamiltonian if
f
v
+
f
e
+ (
n
− 1)
f
s
≤ (
n
− 1)(
k
+ 1) − 2 and
BC
n,k
−
F
is fault-tolerant Hamiltonian-connected if
f
v
+
f
e
+ (
n
− 1)
f
s
≤ (
n
− 1)(
k
+ 1) − 3. To the best of our knowledge, this paper is the first work which takes faulty switches into account to study the fault-tolerant Hamiltonicity and the fault-tolerant Hamiltonian connectivity in data center networks.
The inefficient sharing of industrial cloud resour-ces among multiple users and vulnerabilities of virtual machines (VM)s and servers prompt unauthorized access to users' sensitive data along with ...excess consumption of power and resource wastage. To address these entangled issues, this paper proposes a novel E merging VM T hreat P rediction and Dynamic W orkload E stimation based Resource Allocation ( ETP-WE ) framework that predicts VM threats and resource usage proactively in real-time. The proposed framework contributes by introducing a Risk-Score Matrix that analyses multiple risks for each VM; utilizing knowledge of proposed security and workload analyzers for efficient VM Placement (VMP), and estimating resource utilization by developing an ensemble predictor for prior mitigation of over-/under-load on servers. ETP-WE framework collaborates machine-learning-based security and workload analysis for secure and resource-efficient VMP, thereby reducing the number of security threats, optimizing resource utilization, power-consumption, and adapting to the changes in application demands. The performance of the proposed framework is evaluated using two benchmark datasets OpenNebula and Google Cluster. The simulation-based comparison with state-of-the-arts validates the efficacy of ETP-WE in terms of reduction of security threats, power consumption, and number of active servers up to 86.9%, 66.67% and 30%-80%, respectively with an improved resource utilization up to 60%-75% over existing approaches Note to Practitioners -Industry clouds serve the precise needs and provide the service features and tools as per the industry's needs to help organizations meet their workloads processing and storage demands. For instance, healthcare and financial organizations have to comply with extended security to meet specific service requirements. To this context, we have proposed a novel ETP-WE framework for prediction and mitigation of cyberthreats on virtual resources in real-time for secure execution of industrial applications on third-party servers. ETP-WE collaborates machine-learning based security and workload analysis for secure and resource efficient VM allocation, thereby reducing number of security threats, optimizing resource utilization, power-consumption and adaptating to the changes in application demands. During the processing of any sensitive transaction such as medical data, bank transactions, ETP-WE framework will induce improved data protection by mitigating potential data breaches. ETP-WE framework will be deployed at Resource Scheduler to boost security performance by estimating the multiple risks score status of VMs engaged in execution of industrial transactions or workloads. It will help to predict and analyze the probable security threats or breaches proactively and facilitate their mitigation. The performance evaluation and comparison with state-of-the-arts validate potency of ETP-WE in terms of reduction of cybersecurity threats and number of active servers with an improved resource utilization over existing approaches.
Intrusion detection need grows with the increase in the count and volume of Internet Traffic and Network. In this paper, intrusion detection system (IDS) is proposed to identify and distinguish the ...incoming traffic from the clients and traffic originating through the attackers by using the honeypot security mechanism. Load Balancer is designed and implemented in such a way that it differentiates between the incoming traffic from clients, and the traffic that arises through the attackers. When forwarding the request, it discovers if the traffic is an attack on the server and directs it to a different/alternate server called Honey-Pot. The IDS is developed and intensified with two protocols: File-based detection and Real-time based detection. The File based detection works on question model and the Real-time based detection sets priority levels for accessing data. This acts as a secure-direct protocol thereby increasing the Server protection. The Secure direct method prompts an immediate response to define network intrusions and eliminates human interference to identity the intrusions. Further an interaction with IDS is done to decide whether the traffic is a trespasser. If the traffic is found to be an access user the packet is transmitted to server, but if found to be an unauthorized user, then the IDS directly transmits it to the server.
We study a novel replication mechanism to ensure service continuity against multiple simultaneous server failures. In this mechanism, each item represents a computing task and is replicated into ξ+1 ...servers for some integer ξ≥1, with workloads specified by the amount of required resources. If one or more servers fail, the affected workloads can be redirected to other servers that host replicas associated with the same item, such that the service is not interrupted by the failure of up to ξ servers. This requires that any feasible assignment algorithm must reserve some capacity in each server to accommodate the workload redirected from potential failed servers without overloading, and determining the optimal method for reserving capacity becomes a key issue. Unlike existing algorithms that assume that no two servers share replicas of more than one item, we first formulate capacity reservation for a general arbitrary scenario. Due to the combinatorial nature of this problem, finding the optimal solution is difficult. To this end, we propose a Generalized and Simple Calculating Reserved Capacity (GSCRC) algorithm, with a time complexity only related to the number of items packed in the server. In conjunction with GSCRC, we propose a robust replica packing algorithm with capacity optimization (RobustPack), which aims to minimize the number of servers hosting replicas and tolerate multiple server failures. Through theoretical analysis and experimental evaluations, we show that the RobustPack algorithm can achieve better performance.
Abstract
Advances in mass spectrometry enabled high throughput profiling of lipids but differential analysis and biological interpretation of lipidomics datasets remains challenging. To overcome this ...barrier, we present LipidSuite, an end-to-end differential lipidomics data analysis server. LipidSuite offers a step-by-step workflow for preprocessing, exploration, differential analysis and enrichment analysis of untargeted and targeted lipidomics. Three lipidomics data formats are accepted for upload: mwTab file from Metabolomics Workbench, Skyline CSV Export, and a numerical matrix. Experimental variables to be used in analysis are uploaded in a separate file. Conventional lipid names are automatically parsed to enable lipid class and chain length analyses. Users can interactively explore data, choose subsets based on sample types or lipid classes or characteristics, and conduct univariate, multivariate and unsupervised analyses. For complex experimental designs and clinical cohorts, LipidSuite offers confounding variables adjustment. Finally, data tables and plots can be both interactively viewed or downloaded for publication or reports. Overall, we anticipate this free, user-friendly webserver to facilitate differential lipidomics data analysis and re-analysis, and fully harness biological interpretation from lipidomics datasets. LipidSuite is freely available at http://suite.lipidr.org.
Graphical Abstract
Graphical Abstract
LipidSuite is an integrated webserver for differential lipidomics analysis, supporting interactive data processing and visualisation, as well as lipid set enrichment and chain trend analysis.
Abstract
Analyzing all features of small non-coding RNA sequencing data can be demanding and challenging. To facilitate this process, we developed miRMaster. After the analysis of over 125 000 human ...samples and 1.5 trillion human small RNA reads over 4 years, we present miRMaster 2 with a wide range of updates and new features. We extended our reference data sets so that miRMaster 2 now supports the analysis of eight species (e.g. human, mouse, chicken, dog, cow) and 10 non-coding RNA classes (e.g. microRNAs, piRNAs, tRNAs, rRNAs, circRNAs). We also incorporated new downstream analysis modules such as batch effect analysis or sample embeddings using UMAP, and updated annotation data bases included by default (miRBase, Ensembl, GtRNAdb). To accommodate the increasing popularity of single cell small-RNA sequencing data, we incorporated a module for unique molecular identifier (UMI) processing. Further, the output tables and graphics have been improved based on user feedback and new output formats that emerged in the community are now supported (e.g. miRGFF3). Finally, we integrated differential expression analysis with the miRNA enrichment analysis tool miEAA. miRMaster is freely available at https://www.ccb.uni-saarland.de/mirmaster2.
Graphical Abstract
Graphical Abstract
miRMaster 2 workflow. The workflow is split into three steps: data selection and compression, data upload and server-side processing.