摘要： 在计费系统的日常维护工作中，常遇到由于无法准确预估主机负载而出现的主机资源过度紧张或者闲置。针对该问题，分析计费业务流量与系统负载之间的关系，提出一种基于单位业务tpmC当量的计费系统负载预测模型。对每一类业务计算单笔交易的系统负载，分析并建立系统负载模型。实验结果表明，该模型预测的系统负载数据与实际数据误差在3%以内，有助于提升系统资源的使用效率，保障系统稳定运行。

关键词: 计费系统, 负载模型, tpmC当量, 业务分解, 负载预测, 精确运维

Abstract: Sometimes it is hard to predict the billing system device performance accurately and adjust process resource dynamically, the devices are overloaded under high business pressure and idle when there is no business. Aiming at this problem, this paper analyzes the relationship between business flow and system load, proposes a new modeling method of load model of a billing system, which is based on the tpmC equivalent of every single business process unit, through decomposing business flow of same device and researching the effect of every business flow. Experimental result shows that the error between load data and actual data of predicted system is less than 3%, this model is helpful to enhance the efficiency of system resource, and guarantees the stable operation of the system.

Key words: billing system, load model, tpmC equivalent, service decomposition, load prediction, accurate maintenance

中图分类号: 

• TP311