Zhengdong Zhu

Research on Algorithms to Capture Drivers’ Write Operations

Full virtualization technology is highly reusable. Using this property, various types and versions of existing operating systems and drivers can be reused in a virtual machine to customize users’ application environments. However, these environments are threatened by drivers’ write operation faults, which are caused by bugs in reused drivers. Chariot is a reliability architecture that has been developed to solve this problem. This architecture captures a driver’s write operations by maintaining the write permissions of shadow pages as read-only to examine their correctness. Nevertheless, this capture method produces many page faults in the virtual machine monitor and has an adverse impact on the performance of isolated drivers. To reduce performance losses, this paper examines two algorithms that cache recently used shadow pages using different structures to avoid frequent page faults. The experimental results show that the performance of isolated drivers can be greatly improved using these shadow page caches without significantly impacting the isolation efficiency of Chariot.

Small files storing and computing optimization in Hadoop parallel rendering

The Hadoop framework has been widely used in the animation industry to build a large scale, high performance parallel render system. However, Hadoop Distributed File System (HDFS) and MapReduce programming model are designed to manage large files and suffer performance penalty while rendering and storing small RIB files in rendering system. Therefore, method that merging small RIB files based on two intelligent algorithms is proposed to solve the problem. The method uses Particle Swarm Optimization (PSO) and Support Vector Machine (SVM) to choose the optimal merge value for any scene file, by mainly considering the rendering time, memory limitation and other indicators. Then, the method takes advantage of frame-to-frame coherence to merge RIB files at an interval way with the optimal merge value. Finally, the proposed method is compared with the naive method under three different render scenes. Experimental results show that the proposed method significantly reduces the number of RIB files and render tasks, and improves the storage efficiency and computing efficiency of RIB Files.

Thread Mapping and Parallel Optimization for MIC Heterogeneous Parallel Systems

There is no dedicated thread mapping method for Many Integrated Core (MIC) heterogeneous system in the traditional multithread programming model. The unreasonable thread mapping will lead the promising computing power of MIC coprocessor not to be fully exploited. In order to fully exploit the computing potential of MIC coprocessor, this paper discussed effective multi threads mapping strategies through comparing the computing performance and analyzing the performance differences between various mapping methods. Meanwhile, for the further exploiting the high computing power of MIC heterogeneous system, the specific program porting and performance optimization strategies were explored by using the k-means application program. Experimental results show that the proposed mapping and parallel optimization strategies are effective, which can be guide the programmer to port and optimize applications effectively to MIC heterogeneous parallel system.

Taxonomy and an Ontology for Grid Metrics

Grid metrics is the essential way for both human and software (machine) to understand the state of a grid system and evaluate its performance. To achieve a better insight into the grid system, the knowledge and profiling of grid metrics are indispensable for the selection and calculation of metrics. We formulate the resource model and workflow model in grid, then introduce an ontology to capture critical characteristics of grid metrics. Based on the ontology, a taxonomy is proposed to organize and characterize grid metrics to build general knowledge of them. Better understanding of grid metrics helps for further decision-making such as designing monitoring services or improving system performance.

Thread Count Prediction Model: Dynamically Adjusting Threads for Heterogeneous Many-Core Systems

Determining an appropriate thread count for a multithread application running on a heterogeneous many-core system is crucial for improving computing performance and reducing energy consumption. This paper investigates the interrelation between thread count and computing performance of applications, and designs a prediction model of the optimum thread count on the basis of Amdahls law combined with regression analysis theory to improve computing performance and reduce energy consumption. The prediction model can estimate the optimum tread count relying on the program running behaviors and the architecture characteristics of heterogeneous many-core system. Using the estimated optimum thread count, the number of the active hardware threads and processing cores on the many-core processor is dynamically adjusted in the process of thread mapping to improve the energy efficiency of entire heterogeneous many-core system. The experimental results show that, using this paper proposed thread count prediction model, on an average, the computing performance is improved by 48.6%, energy consumption is reduced by 59%, and additional overhead introduced is 2.03% compared with that of the traditional thread mapping for the PARSEC benchmark programs run on an Intel MIC heterogeneous many-core system.

Automated Trust Negotiation Based on Concurrent Zero-Knowledge for e-Business Applications

Exchange of attribute certificates is a means to establish mutual trust between strangers wishing to share resources or conduct business transactions. Automated trust negotiation (ATN) is a promising approach to regulating the exchange of sensitive information during this process. It has been a fundamental but challenging problem to preserve the privacy of the two negotiation parties during the period of ATN. We present the enhanced hidden credentials and improved concurrent zero-knowledge proof protocol. Based on the above technologies, we propose an ATN for e-business applications, named CASTLE. CASTLE can not only enable the oblivious and selective usage of an attribute or a certificate, but also be resistible for many attacks, especially conspiracy attack. We illustrate the usage of CASTLE through a typical example.

CASTTE: A Trust Management for Securing the Grid

It has been a fundamental but challenging problem to gain assurance of the trustworthiness of service providers or requesters and ensure their interests. We present the formal definition of trust management, and then propose a trust management, CASTTE, to secure sensitive services and requesters in grids. CASTTE verifies access trust by using trust negotiation so as to protect sensitive services, and protects sensitive information of the two negotiators effectively by using a negotiation strategy based on protection tree. Furthermore, we utilize trust force to specify provision trust and apply trust force to service selection. This paper implements CASTTE and designs experiments to evaluate its performance. The experimental results show that it can not only protect sensitive services at the cost of little performance of systems, but also identify good services from bad ones effectively.

DOMe: A deduplication optimization method for the NewSQL database backups

Reducing duplicated data of database backups is an important application scenario for data deduplication technology. NewSQL is an emerging database system and is now being used more and more widely. NewSQL systems need to improve data reliability by periodically backing up in-memory data, resulting in a lot of duplicated data. The traditional deduplication method is not optimized for the NewSQL server system and cannot take full advantage of hardware resources to optimize deduplication performance. A recent research pointed out that the future NewSQL server will have thousands of CPU cores, large DRAM and huge NVRAM. Therefore, how to utilize these hardware resources to optimize the performance of data deduplication is an important issue. To solve this problem, we propose a deduplication optimization method (DOMe) for NewSQL system backup. To take advantage of the large number of CPU cores in the NewSQL server to optimize deduplication performance, DOMe parallelizes the deduplication method based on the fork-join framework. The fingerprint index, which is the key data structure in the deduplication process, is implemented as pure in-memory hash table, which makes full use of the large DRAM in NewSQL system, eliminating the performance bottleneck problem of fingerprint index existing in traditional deduplication method. The H-store is used as a typical NewSQL database system to implement DOMe method. DOMe is experimentally analyzed by two representative backup data. The experimental results show that: 1) DOMe can reduce the duplicated NewSQL backup data. 2) DOMe significantly improves deduplication performance by parallelizing CDC algorithms. In the case of the theoretical speedup ratio of the server is 20.8, the speedup ratio of DOMe can achieve up to 18; 3) DOMe improved the deduplication throughput by 1.5 times through the pure in-memory index optimization method.

Dynamic Token Based Improving MapReduce Performance in Cloud Computing

In recent years, Hadoop, the open-source implementation of Googles MapReduce, is widely used and has become the de facto standard of big data processing. A typical running environment of Hadoop is cloud computing in which resource heterogeneity is very common due to varied factors including the different hardware of nodes and the different workload on the nodes and etc. The slot-based scheduling in Hadoop causes the inefficient utilization of computing resources in cloud computing environment, which lead to the performance degradation. To solve the problem mentioned above, we propose a dynamic token based method which dynamically controls the number of tasks running on each node according to the available computing resources on a node and the resource requirement of a task. The results of evaluations show that the completion times of single jobs with the proposed method are approaching to the static optimum in the dedicated environment and better than the static optimums in the other two competitive environments. Moreover, the proposed method significantly improves the throughput of mixed workloads in all computing environments and performance in real cloud computing environment have been improved by 45.9% on average.

Parallel Optimization Strategies for MIC Heterogeneous Parallel Systems

In the traditional multithread programming model, there is no dedicated performance optimization strategy for Many Integrated Core (MIC) heterogeneous system. To fully exploit the high computing power of MIC processor, this paper discusses the specific program porting and performance optimization strategies on the MIC heterogeneous parallel system based on the k-means application program. Experimental results show that the proposed porting and performance optimization strategies are effective, and can be able to guide the programmer to port and optimize applications effectively to MIC heterogeneous parallel system.