农业生产及环境的信息采集是农业信息系统的重要环节，无线传感器网络以它的低成本、无线传输等优势广泛应用于农产品安全追溯的信息采集中。由于无线传感器网络长年不间断信息采集的积累，数据库的大信息量及网站用户高访问量将对网站服务器造成高负载的压力，这将导致网站的响应速度大大降低。采用高存储、高负载处理性能云计算计算技术，能提高网站访问的响应速度，提高网站的访问性能。该文设计了基于私有云的农产品安全追溯网站，研究了Hill-Climbing搜索算法优化云平台配置及MapReduce对大数据的并行计算，提高了私有云上的农产品安全追溯系统的性能。通过将网站系统迁移至私有云的前后的比较试验，说明将网站系统迁移至私有云后各方面性能得到较大提升，网站访问响应速度提高了33%。%Information collection of the production, supply and marketing chain in agriculture is an absolutely necessary part in the driving process of construction of agricultural informationization. As a new technique in collecting information, and relying on its low cost, low power consumption and wireless transmission, a Wireless Sensor Network (WSN) has already permeated into the agriculture area. For an agricultural information system which employs a wireless sensor as its collection terminal, there is no doubt that the uninterrupted process for information collection, transmission, and storage, as well as inquiries from a large number of users, puts so much pressure on the system server, resulting in a low speed for web access and inquiry. The response time and response rate when visiting a web site are key points of a web site’s performance. A supporting platform that has the capacity of mass storage and dealing with high concurrency traffic, therefore, is necessary for a Safety Tracing System for Agricultural Products. A cloud computing platform has an ability of cooperative working with multiple servers, and is able to share the load pressure caused by dealing with mass storage and high concurrency traffic. Following the agricultural Internet of Things, it is necessary to introduce a cloud computing platform to agriculture. The research group has undertaken 13 projects assigned from Science and Technology Department of Guangdong along with Economy and Information Commission in the past two years, including breeding safety, tracing, and electronic business for agricultural products. In order to meet the demand from the Economy and Information Commission, different platforms will be combined into one-Safety Tracing and Trade Platform for Agricultural Products. Recently, the research team has paid attention to setting up the Safety Tracing Platform on the basis of the breeding Internet of Things, dealing with some agricultural products like chickens, rabbits, fish, flocks, and herds. As a result of there being so many kinds of products and mass data required, it is quite necessary to bring a cloud computing platform into this system. It can raise the response rate of visiting a web site and improve the visiting performance by using a cloud computing technique that has a good performance on high storage and heavy load processing. A Safety Tracing System for Agricultural Products based on a cloud computing platform has been designed, and the performance of a system web site has been promoted. Based on a cloud computing platform, the Safety Tracing System for Agricultural Products described in this paper can be divided into three sub-systems logically - a cloud computing platform, a Safety Tracing Subsystem for Agricultural Products, and an Information Collection Subsystem in a WSN environment which is installed in a Zhuhai Production Base. The Bingo cloud platform based on Amazon EC2 was used to develop a private cloud in a cloud computing platform. The SQL Server database and NET language were used to develop a Safety Tracing Subsystem for Agricultural Products. WSN was adopted in an Information Collection Subsystem to collect indicators like temperature, PH value, and humidity during the production. As the first step of the project, the cloud computing platform was constructed, the Safety Tracing Subsystem for Agricultural Products was put onto the platform, and the Wireless Sensor Acquisition Subsystem was connected with the platform, and the construction of the Safety Tracing Subsystem for Agricultural Products based on a cloud computing platform has been finished so far. Secondly, a search algorithm called Hill-Climbing was used to optimize the Safety Tracing Subsystem for Agricultural Products based on a cloud computing platform. Then, to deal with mass data, Mapreduce was used for concurrent processing. The map function receives data and outputs key-value pairs as the middle output after processing. The output from the map function was processed by the Mapreduce framework before being sent to the reduce function. This processing sorts and groups the key-value pairs by keys. Then the reduce functions receive the pairs with the same key and give the final output. This is the advantage of cloud computing:a mass data is cut up into pieces and are passed to CPUs for concurrent processing, which improves the performance of the system on the cloud platform. Here is how the system works:wireless sensor transmits the data of the field information collected from a locale to the cloud database. The cloud computing platform will automatically assign the storage to the data. When customers trace the product information, the platform will automatically balance the tasks of information querying with high traffic and heavy load, making several servers working cooperatively so that it can return the information to customers quickly. Besides, the platform also can use Mapreduce to cut up a mass data into pieces for many CPUs to process concurrently. Research on the heavy load processing performance of cloud computing platform when dealing with storing mass data and high traffic was conducted and the response time of dealing with high traffic based on the platform were described in this paper. The capacity of processing by sharing the heavy load also examined. By comparing the capacity of the cloud computing platform server with that of a general server, the results indicated that there is an all around advance after carrying the system from a general server to a cloud computing server. The response rate of visiting a web site was raised by 33%.
展开▼
