Huawei collaborates with some Chinese large busi-ness companies to store and process exabytes of nationwide operational data in data lake storage to provide business insights. Specifically, our customers will ask to store and process massive log message data to support their real-time and decision-making applications. Thus, we need computation and storage components in the analytic platform to process and store these data cost-efficiently. To meet these user requirements, we have designed a storage system in data lake, StreamLake, which introduces a novel design to serve log message streaming and batch data processing in distributed storage, with high scalability, efficiency, reliability and low cost. Specifically, we introduce a stream (storage) object as a storage abstraction for message streaming data to achieve the storage-disaggregated architecture with high scalability and reliability. Moreover, we utilize the erasure coding and tiered storage to save the storage cost, and furthermore, the stream object can be automatically converted to a table object such that cost-effective stream and batch data processing can be achieved. For tabular data, we implement the lakehouse functionality to support ACID via the table object, with a metadata acceleration to improve the efficiency of data access between the compute and storage engines. Also, we design a LakeBrain optimizer at the storage side to optimize the query performance and resource utilization under the storage-disaggregated architecture. Finally, we have also deployed StreamLake in China Mobile, the world's largest mobile network operator to serve over 20PB production data, and the results demonstrate improvements of 30% to 4x in terms of query performance and over 37% in terms of cost saving.

HAWQ, developed at Pivotal, is a massively parallel processing SQL engine sitting on top of HDFS. As a hybrid of MPP database and Hadoop, it inherits the merits from both parties. It adopts a layered architecture and relies on the distributed file system for data replication and fault tolerance. In addition, it is standard SQL compliant, and unlike other SQL engines on Hadoop, it is fully transactional. This paper presents the novel design of HAWQ, including query processing, the scalable software interconnect based on UDP protocol, transaction management, fault tolerance, read optimized storage, the extensible framework for supporting various popular Hadoop based data stores and formats, and various optimization choices we considered to enhance the query performance. The extensive performance study shows that HAWQ is about 40x faster than Stinger, which is reported 35x-45x faster than the original Hive.

Choosing suitable and excavating potential civil aircraft suppliers are key links of realizing the aircraft collaborative development, and important factors to improve competitiveness of the aviation supply chain. In view of this, according to the characteristics of civil aircraft supplier selection, imitating the idea of multi-target weighted grey target decision and using the grey clustering method to construct the airliner supplier evaluation clustering method. Firstly, indicator system of civil aircraft supplier evaluation is designed, and the multi-objective intelligent weighted grey target decision method is used to dimensionless processing for supplier evaluation indicators, and then the improved triangle whitening weight function clustering method is applied to evaluating suppliers clustering, lastly the validity and rationality of the model is verified by giving an example of supplier selection for an aircraft structure part clustering evaluation. The results show that the method constructed can handle aircraft supplier evaluation and clustering problems well, and choose aircraft supplier effectively.

Knowing accurate positions of nodes in wireless ad hoc and sensor networks is essential for a wide range of pervasive and mobile applications. However, errors are inevitable in distance measurements and we observe that a small number of outliers can degrade localization accuracy drastically. To deal with noisy and outlier ranging results, triangle inequality, is often employed in existing approaches. Our study shows that triangle inequality has many limitations, which make it far from accurate and reliable. In this study, we formally define the outlier detection problem for network localization and build a theoretical foundation to identify outliers based on graph embeddability and rigidity theory. Our analysis shows that the redundancy of distance measurements plays an important role. We then design a bilateration generic cycles-based outlier detection algorithm, and examine its effectiveness and efficiency through a network prototype implementation of MicaZ motes as well as extensive simulations. The results show that our design significantly improves the localization accuracy by wisely rejecting outliers.

This study investigates the role of psychological empowerment in the customer evaluation. To test effect, employees and customers were surveyed in 4S stores. We aggregated the employee-individual and customer-individual level data to a higher unit level, and a gray preference model was proposed to explain the role of empowerment in customer evaluation. This study demonstrates that customer loyalty is more favorable than customer satisfaction as for system characteristic behavior, and managers need to focus on employee competence to improve customer evaluation.

Hadoop Distributed File System (HDFS) has been widely adopted to support Internet applications because of its reliable, scalable and low-cost storage capability. Blue Sky, one of the most popular e-Learning resource sharing systems in China, is utilizing HDFS to store massive courseware. However, due to the inefficient access mechanism of HDFS, access latency of reading files from HDFS significantly impacts the performance of processing user requests. This paper introduces a two-level correlation based file prefetching approach, taking the characteristics of HDFS into consideration, to improve performance by reducing access latency. Four placement patterns to store prefetched data are presented, with policies to achieve trade-off between performance and efficiency of HDFS prefetching. Moreover, a dynamic replica selection algorithm is investigated to improve the efficiency of HDFS prefetching. The proposed prefetching approach has been implemented in Blue Sky, and experimental results prove that correlation based file prefetching can significantly reduce access latency therefore improve performance of Hadoop-based Internet applications.

Knowing accurate positions of nodes in wireless ad-hoc and sensor networks is essential for a wide range of pervasive and mobile applications. However, errors are inevitable in distance measurements and we observe that a small number of outliers can degrade localization accuracy drastically. To deal with noisy and outlier ranging results, triangle inequality is often employed in existing approaches. Our study shows that triangle inequality has a lot of limitations which make it far from accurate and reliable. In this study, we formally define the outlier detection problem for network localization and build a theoretical foundation to identify outliers based on graph embeddability and rigidity theory. Our analysis shows that the redundancy of distance measurements plays an important role. We then design a bilateration generic cycles based outlier detection algorithm, and examine its effectiveness and efficiency through a network prototype implementation of MicaZ motes as well as extensive simulations. The results shows that our design significantly improves the localization accuracy by wisely rejecting outliers.

The post evaluation of construction projects is an important link of the whole project management. The indexes of post evaluation usually include preference information. In the paper, dominance-based rough set and grey relational analysis (GRA) are combined to apply to the process evaluation in the post evaluation of construction projects, and a formal and effective method of post evaluation is proposed. Preferential decision models are induced from the decision tables with preferential information. The effectiveness of the results demonstrates that the inferring preference model is close to the natural reasoning of decision makers, and is comprehensible. The preferential rules derived from examples can provide decision suggestion for the post evaluation of the construction project. Moreover, the relative importance for reduced attributes is wholly compared by grey relational degree, which provide basis with further analysis, such as setting weight and making mathematic models.

The degree of uncertainty to which an object belongs to a class or not being viewed as grey degree of grey numbers, an approach to grey measure of knowledge is proposed based on rough membership function, which provides a formal reasonable definition of grey degree of grey numbers and grade of grey numbers. It is shown that the semantics of grey degree of grey numbers can be explicitly stated, the levels of grey numbers can be reasonably interpreted for objects to classify and the set of rules can be obtained with different grades of grey degree.