Font Size: a A A

The Research Of AST3 Realtime Computing System

Posted on:2011-06-08Degree:MasterType:Thesis
Country:ChinaCandidate:W CaoFull Text:PDF
GTID:2178330338989194Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
AST3 Project is an important part of Chinese Antarctic Expedition. The plan is to deploy 3 Schmidt telescopes in the Antarctic Kunlun Station, which used to do the unattended observation and real-time data processing in the Antarctic polar night. The first telescope and its data processing computer will be transported to Kunlun Station during the 27th Antarctic Expedition and put into use in 2011. As the unique conditions of Antarctic's geographical location and climate, it is not yet possible to do the observation with any manual control, so the data processing must be triggered and managed automatically; the same time, the 4-months observation will produce a large amount of data, only a little of which can be sent back by satellite for remote processing, thus requiring AST3 system has capabilities of real-time data processing and analysis in local. So a high quanlity software system must be designed in order to meet the reliability under no duty, as well as real-time data processing requirements.Our research focous on providing a solution for the AST3 Project. In this paper, we design and implement such a software system, which is based the mechanism of Linux daemon, we called AST3 Daemon service. By this service, it is easy to integrate all kinds of data processings, and provide the system-wise reliability and robustness. With the integration of "autonomic computing" module, we provide some reliability assurance mechanisms such as crash recovery, adaptive and automatic decision-making. General GPU-based computing, OpenMP and other parallel computing based methods are used to accelerate data processing to meet the real-time requirements. A complete set of iterative testing programs, including module unit tests and integration tests, are designed for all phases of the software life cycle, in order to meet the stringent requirements of software quality.
Keywords/Search Tags:Astronomy Image Processing, Parallel Computing, General Purpose GPU Computing, Autonomic Computing
PDF Full Text Request
Related items