Reseaarch On Cloud-based Storage Information System

Cloud computing is a shared network mode for delivering information service. Cloud service users see only the service itself, rather than the concrete realization of the related infrastructure. Two essential characteristics of cloud computing are resource virtualization and computing parallelization. Resource virtualization is the logical representation of computing resources and storage resources. It is not subject to physical limitations.The main objective of the existing cloud storage architecture is to address the problem of how to make a reasonable abstraction of cloud computing and cloud storage in the Internet environment, and then to establish a cloud storage system architecture model which reflects the main features of the new generation of storage systems. These researches have this in common which regards the storage media as the smallest units of cloud storage. The cloud storage is not able to form the cloud parallel processing architecture of the storage media, for it does not penetrate the smallest units.This thesis focuses on the research of the structure for the storage media and proposes a new type of storage structure:cloud-based storage information system (CBSIS). CBSIS distributes file management systems and storage media into parallel processing of the cloud architecture. That is to map file management system and storage media to the atomic component (AC) and PRM engine (E) on the basis of establishing CBSIS meta-model or rather Problem Solving Model (PRM); to realize the AC of the atomic system components (AS) by using the functional integrated circuit IP cores; and to make AS into the PRM parallel tree leaf nodes. The visits among PRM tree nodes are message-based. Ultimately all these above contribute to the formation of cloud storage architecture of the system on a chip (SSoC)This doctoral thesis focuses on three innovative research results from four aspects. That is, firstly studies storage system architecture model based on cloud computing architecture according to the two essential characteristics of resource virtualization and cloud computing in parallel; this thesis then analyzes, designs and validates new generation CBSIS cloud storage system based on this architecture model; and finally a physical prototype system of cloud-based storage system on-chip has been implemented.(1) The thesis researcher has established the cloud storage system architecture model CBSIS based on the PRM meta-model which is the first originality of the research. This research has established the problem-resolving model for specific issues, which includes a conceptual model, logical and physical model of the three-layer SOA architecture. The physical model is mapped to SOA components, which makes a limited set of atomic components of the component pool by way of the logic model to perform the semantics of infinite set of the conceptual model. The relationships of these components are loosely coupled, and visits and access of these components are message-based through the interface and virtualization basis of the cloud computing resource.(2) The overall framework based on CBSIS architecture is characterized by:(a) the introduction of architecture into CBSIS system design;(b) the connection of the engine E with the architecture connectors CON so that the PRM deployment and the CON deployment are consistent. The second innovation of this research is to propose and define the CBSIS architecture description tool LL7, and to implement LL7by building the connections of the CON model and the AC model.(3) Having taken the generating mode of CBSIS executable prototype system as the background, this researcher has made some discussions on generating executable prototype business system and executable architecture prototype system constraints, and has defined the concept of incremental prototype system. The researcher also points out that the advantages of the incremental prototype system lie in analysis, design, and implementation stages.(4) The proposed solution based on the system SSoC cloud storage chip is characterized with cloud storage system status, future development trends, the phased transformation of cloud storage infrastructure management platform, and the entire data center deployment and implementation. Taking the cloud storage infrastructure improvements to the cloud storage architecture as an example, the thesis proposes and implements SSoC based on the SoW and bSATA controller, which forms the third outcome of this research. The SSoC is characterized by the integrated File Management System model and the hierarchical nested PRM tree structure formed by the model elements.Although this research focuses cloud storage applications, the methods and tools mentioned here in the thesis can be widely used elsewhere. Therefore, the proposed methods and philosophy of this research is applicable to the general application of systems modeling, with characteristics of rapid development and deployment, ease of testing and ease of maintenance.