以分布式计算实现电信数据分析业务加速的研究

doi:10.3969/j.issn.1000-0801.2012.02.005

Abstract

Abstract:

The paper analyzes the problems of data analysis system of telecommunications.With the business needs and technology trends,it proposes a solution based on Hadoop distributed computing framework to accelerate data analysis for the telecommunications infrastructure services.Then,this paper describes a framework including technology solutions,functional model,and acceleration which based on distributed data analysis system.And also does a further analysis and comparison with the application of distributed computing.This paper will provide a useful reference for the industry of distributed computing applications in related fields.

Key words: distributed computing, mass data, business intelligence, IDAP

Jianfeng Wen,Yirong Shi. Research on Basing Distributed Computing Framework to Achieve Accelerating the Data Analysis Application of Telecommunications[J]. Telecommunications Science, 2012, 28(2): 22-26.

取数业务	取数主题	取数用例	传统取数时间（s）	分布式取数时间（s）
收入类取数	设备级收入取数（1个月）	20110914-141049-tongj	1 355	39
		20110927-111311-wj31027332	2 635	355
	设备级收入取数（3个月）	20111010-164109-fanjie1	4 089	237
		20111009-232626-xud31099100	2 155	452
	客户级收入取数（1个月）	20110718-151855-maxiaoyuan	785	41
	客户级收入取数（3个月）	20110727-180632-qihua	349	96
	财务收入明细清单（区局）	20111011-094037-yuanfenghua	662	161
订单类取数	订单根行项目完工清单	20111011-174303-yuanfenghua	90	199
	订单子行项目完工清单	20110915-115542-huyizhen	1 702	132
	订单根行项目受理清单	20110915-143605-taohui	113	70
	订单子行项目受理清单	20110913-140837-wengzhenhua	650	117
资产类取数	根资产当前信息	20111011-104713-zhaohaigang	2 181	147
	根资产历史信息	20111006-223219-xud31099100	840	127
	子资产历史信息	20110913-093139-yaojie	47 583	786
业务量取数	设备级固话业务量取数（1个月）	20110906-125327-cl31800892	4 241	430
		20110801-093025-mafang	1 177	166
	设备级CDMA网业务量取数（1个月）	20111011-163729-huyizhen	1 105	106
	设备级宽带业务量取数（1个月）	20110913-155017-wengzhenhua	1 824	186
	设备级固话业务量取数（3个月）	20110809-104344-maxiaoyuan	3 617	280
		20111013-010932-xud31099100	8 848	653
	设备级CDMA网业务量取数（3个月）	20110801-094619-qihua	6 486	471
		20110905-182616-xud31099100	2 263	376

	传统方案	分布式方案一	分布式方案二
技术方案	Oracle	Hadoop	Hadoop
查询方法	数据宽表SQL查询	数据宽表SQL查询	数据宽表SQL查询
数据接口	JDBC	JDBC	JDBC
硬件平台	小型机+SAN存储	20台工控机	5台PC服务器
硬件成本	100万元	14万元	14万元
软件成本	Oracle许可费	免费	免费
扩展成本	实施成本+许可	实施成本	实施成本
加速效果	-	10.92倍	6.95倍
空间利用率	2个机柜	未节约	1/2个机柜，节约75%
能耗	10 000 W	20×300 W，节约40%	5×600 W，节约70%

[1]	Ping XIE,Xiaosong LIU. Security of the deployment of SDN-based IoT using blockchain [J]. Telecommunications Science, 2020, 36(12): 139-146.
[2]	Jianxin Ren,Huahui Chen. An Adaptive Subspace Similarity Search Approach [J]. Telecommunications Science, 2015, 31(7): 63-74.
[3]	Jie Zheng,Sheng Bao,Ying Yang. Application of Distributed Parallel Computing in Super-Resolution Image Enbancement [J]. Telecommunications Science, 2015, 31(4): 37-42.
[4]	Jie Liu,Zhe Wang. Construction of Carrier Business Precise Operating Platform Basedon Big Data [J]. Telecommunications Science, 2013, 29(3): 22-26.
[5]	Julong Lan,Dongnian Cheng,Yu Wang,Fengyu Zhang. Initial Analysis on Intelligence Mechanisms of Reconfigurable Network Abstract Flexibility,with which a reconfigurable network adaptively tunes its structure,plays a critical role in accommodating varieties of applications with desirable service levels.Most studies avalaible,however,rarely provide clear answers about the mechanism of flexibility that supports network reconfigurations.For the mechanism more clear,this paper clarifies a key target of the consistant matching of service outcomes to application demands.What is then revealed includes the kernal characteristic,so-called outcome following demand with gardually tuning,of the reconfiguration flexibility.Furthermore,four inherent schemes and frameworks representing reconfiguration flexibility are respectively examined in depth.They are the knowledeg-driven architecture of network functionalities,cross-layer interactions of functionalities,distributed and collaborative sensing,and distributed intelligece.Finally,the method of ant conoly optimization（ACO）is emphasized by explaining its effectiveness in optimizing non-linear problems such as performence-constrained routing at small time scales.Specifically,both the advantage and significance of emergence,the prominent feature of ACO,are illustrated. [J]. Telecommunications Science, 2012, 28(8): 105-112.
[6]	Xingming Guo,Yugang Zhu,Kaijun CHEN,Yong He. Business Intelligence Ontology Cloud Architecture Design [J]. Telecommunications Science, 2011, 27(11): 35-40.
[7]	Yang Zhang. Business Intelligence as a Service—SaaS BI Introduction and Recommendations [J]. Telecommunications Science, 2010, 26(8A): 85-88.
[8]	Dezhi Gong,Yong Duan. Application Research in Cloud Computing of Data Analysis and Business Intelligence Analysis for Telecommunication [J]. Telecommunications Science, 2010, 26(6): 30-35.
[9]	Huayun Yan,Jihong Guan. Survey of Bloom Filter [J]. Telecommunications Science, 2010, 26(2): 31-36.

Research on Basing Distributed Computing Framework to Achieve Accelerating the Data Analysis Application of Telecommunications

RichHTML

PDF下载

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 4

References 3

Related Articles 9

Metrics

Recommended 0

1	闻剑峰，石屹嵘 . 基于云计算的全球眼业务平台研究. 电信科学， 2010,26（6）
2	龚德志，闻剑峰 . 虚拟化技术在电信服务器资源整合中的应用研究. 电信科学， 2009,25（9）:21～23
3	石屹嵘，段勇 . 云计算在电信IT领域的应用探讨. 电信科学， 2009,25（9）:24～28