- 學位論文
平行化處理資料倉儲系統產業之策略研究
A Study on the Strategy of the Massive Parallel Data Warehousing System Industry
摘要
資料倉儲系統 (Data Warehousing System) 已是現代企業中的重要資訊系統之一，其主要價值在於增進企業決策之決策品質，並具有增強企業競爭力的策略性意義。非平行化資料倉儲系統底層為關聯式資料庫，其IT基礎架構(伺服器、光纖網路交換器、磁碟陣列)對於數個Terabytes以上的大量資料，並無法有效處理。隨著企業內營運與決策所需資料之逐年增加，平行化資料倉儲系統之重要性將與日俱增。台灣企業採用平行化資料倉儲系統已有近20年歷史，但迄今僅限各產業頂端企業，幾乎為Teradata一家廠商所獨佔，且其處理之資料量多數僅在數個Terabytes以下，並未能有效發揮平行化資料倉儲系統之潛力。本研究之目的，在於分析平行化資料倉儲系統產業之廠商策略，以分析除了Teradata之外平行化資料倉儲系統廠商可採取之策略。同時，本研究也欲藉此檢視資源基礎理論的實用性。 本研究首先介紹平行化資料倉儲系統之特性與技術架構，作為價值鏈分析之基礎，並介紹替代性設計及其特性，以作為五力分析替代品威脅分析之基礎。其次整理市場規模與市佔率之資訊，並逐一分析產業內各主要廠商之規模、歷史、產品特性等，作為產業分析之基礎。本研究接者以價值鏈分析、資源基礎分析、五力分析、與價值網分析等模式，詳細分析本產業內廠商可採取之策略。最後，本研究分析台灣企業採用平行化資料倉儲系統之現況，並將前述之策略分析結果，依據台灣市場現況予以修正。 本研究認為，平行化資料倉儲系統必須與資料分析軟體進行整合，方能展現其價值。因此，本產業廠商必須與資料分析軟體廠商、行業別軟體廠商等合作，整合雙方產品並針對效能予以最佳化，平行化資料倉儲系統之價值才能明顯呈現。此外，由於平行化處理能力為本產業之稀有關鍵性資源，廠商必須提高服務等級，協助客戶將其應用系統之效能調整至明顯優於非平行化系統，否則合作夥伴在並未具有此一關鍵性資源之下，無法有效展現平行化資料倉儲系統之價值。另一方面，MapReduce雲端運算為本產業之替代性技術，其潛在威脅性高。廠商應充分利用雲端運算之趨勢，積極引入此一技術並與傳統平行化資料倉儲系統技術整合並區分其使用範疇，藉以提高本身之差異化並降低產業結構變化時之風險。 在台灣市場方面，本產業廠商宜以製造業大廠之工程分析為首要的市場切入點，成立區域性服務組織，支援台灣之系統整合商與分析軟體代理商，並與互補性廠商合作，自需求面以行銷方式宣導大量資料處理之效益，以扭轉台灣廠商習於以平行化資料倉儲系統處理少量資料之傳統作法。 在理論方面，本研究認為資源基礎理論在策略管理內部分析上，確可作為價值鏈分析的互補性分析架構，但在系統性辨識關鍵性資源、價值性創造的量化分析上，尚需學術界的努力。
並列摘要
Data Warehousing System has been an important information system of modern enterprises. Its major value is to improve the quality of business decisions, and it has the strategic potential of strengthening the competitive advantages of an enterprise. The basis of non-massive parallel processing data warehousing systems is relational database, but its common IT infrastructure, i.e. servers, storage area network switches, and disk arrays, fails to process effectively data volume larger than a few terabytes. As the trend of ever expanding data volume for operational and strategic decisions of enterprises is unavoidable, the importance of massive parallel processing (MPP) data warehousing system increases as well. It has been almost 20 years since the first MPP data warehousing system was installed in Taiwan, but its usage is restricted to top enterprises of each industry only, monopolized by one vendor Teradata, and the data volume handled by such systems is mostly in the rage of 5-10 Terabytres, hardly the true potential of MPP data warehousing systems. The purpose of this research is to analyze the plausible strategies for the companies other than Teradata in this industry. In addition, it is intended to analyze the applicability of resource-based view theory as an analysis framework for strategic analysis. The thesis first introduces the technical architecture of MPP data warehousing systems, as a basis for a value chain analysis, and then describes alternative designs, as a basis for analyzing the threat of substitutes within the five foce framework. Next, the thesis describe industry’s major players, including market share, financial conditions, history, product characteristics, etc., as a basis for an industry analysis. Based on the information, the theis conducts strategy analysis based on four frameworks: value chain, resource-based view, five force framework, and value net. Finanaly, the thesis provides Taiwan market information, and fine tune the strategy analysis results based on local market conditions. The research suggests that an MPP data warehousing system should integrate with data analysis software in order to provide maximum values to the customer. Therefore, vendors in this industry are advised to cooperate with data analysis solution vendors, vertical application vendors, etc., integrate these products and optimize the performance. In addition, as massive parallel processing capabilities is a rare, critical resource of this industry, the vendors should make effort to raise its service level in order to help the customers improve their applications’ performance and demonstrate the values of MPP data warehousing systems. Otherwise, without such critical resources, partner companies such as SI companies are not capable enough to bring the values to the customers. On the other hand, MapReduce-based cloud computing technology is a substitute to the products of this industry, having high potential threats. The MPP data warehousing vendors are advised to take advantage of the cloud computing trend, aggressively adopt the cloud technology, integrate it with the MPP data warehousing technology, and make effort to identify market segments for each of the two technologies, so as to differentiate the vendor itself among others, and reduce the risk of major industry restructuring. In terms of the Taiwan market, the vendors of this industry is advised to target the engineering data analysis applications of major manufacturers, as an entry point for the Taiwan market, organize regional service organizations to support the system integrators and data analysis agent in Taiwan, and work with complementary companies to advertise to the enterprises the benefit of processing large volume of data, in an effort to change the behavior of the Taiwanese enterprises that processing small volume of data using MPP data warehousing systems. The research confirms that resource-based view is a valuable framework for internal analysis of strategy management, as a complementary framework of value chain analysis. However, the theory lacks of a systematic method to identify critical resources, as well as a quantitative approach for analyzing the value creation of resources.