以賽局理論分析台灣Unix伺服器廠商的競合策略

Abstract

伺服器市場在全球的總產值在2006年已經達到523億美元，其中Unix伺服器佔了172億美元(約占了32.9%伺服器市場總額)，幾乎大部份中、大型企業在運行資訊中心時所倚賴的運作核心都採用這種伺服器。在這個市場領域中僅由四家IBM(市佔率為32.8%)、Sun(市佔率為30.3%)、HP(市佔率為27.5%)及Fujitsu(市佔率為4.6%)具代表性的廠商扮演主要的生產及銷售的角色。在這個寡占競爭的市場中經過了近二十年的發展歷史，從作業系統的發展、微處理器的設計研發生產製造、再到整體的主要的零組件幾乎都是處在封閉的環境中運作。這個產業的不透明性也隨之增高。 基於全球Unix伺服器廠商間的競合策略下，台灣Unix伺服器市場各廠商之間的競合策略有何變動，有甚麼重要的因素會影響廠商間的競合策略，本研究藉由多輪出價談判賽局的不合作賽局為基礎，從納許談判解合作賽局的架構中找出Sun及Fujitsu在台灣Unix伺服器市場合作的均衡解，但由於雙方長期競爭的情況下，在實際的談判過程中，雙方對互相所有的威脅點的認知未必相同，因而造成威脅點的不確定性，本研究也以可變威脅的談判賽局來分析雙方要如何操作談判協議中的最佳選擇以符合雙方利益。並依據謝林的「邊緣運用策略」，「可信的承諾」，「有效的嚇阻」等方式提出雙方談判策略的建議。 待雙方達成協議後，在台灣的Unix伺服器市場就形成三強鼎立的局面，本研究再以完美訊息擴展型式賽局中，輪流出招來分析三者之間的競合關係。並再比較同時出招與輪流出招的異同後，提出三方的最佳競合策略。但不論從三方輪流出招或是同時出招的賽局都必須考量： (1). 比較弱的兩方(Solaris及HP)的優勢策略是都要向IBM出招以求得最大的市場利益； (2). 隔山觀虎鬥的策略是不太適合三方對峙的賽局。Solaris陣營可能無法袖手旁觀，任由IBM及HP互相競爭，而不參與其中。 本研究希望透過賽局理論的架構及推論的方法，為產業界以科學的方法帶來新的思維及策略性思考，也期望本研究能幫助產業界從零合賽局的競爭到非零合賽局的合作以爭取更大的超額報酬(extra pay-off)。

The server market has already reached USD 52.3B in 2006 in the total production value of world, among them, the Unix server had USD 17.2B(occupied 32.9% server market a total amount around), almost and most in, large and mid-range enterprise while circulating information center rely on operation cores all adopt this kind of server. It is only played main production and the role of sale by four vendors, IBM(occupied the market share is around 32.8%), Sun(occupied the market share is around 30.3%), HP(occupied the market share is around 27.5%) and the Fujitsu(occupied the market share is around 4.6%). These representative manufacturers are acting main production and the sales role in this market. Widowed occupied the competition in this in the market to pass through the near 20 year development history, from the operating system development, the microprocessor design research and development production manufacture to the main component and parts, nearly all was again occupies in the closing environment e. It is result in this industrial opacity also advances along with it. Competes based on the global Unix server vendor between gathers under the Co-Opetition strategy. Between the Taiwan Unix server market various vendors competes gathers the Co-Opetition strategy to have what change, what important factor has to be able to affect between the vendor to compete gathers the Co-Opetition strategy. This research based on alternating-offers game which the game belong to non-cooperative game to build the analyze structure. And through the architect of Nash-Bargaining Solutions Cooperative Game to find out the Nash-Equilibrium of the cooperative game of Sun and Fujitsu in Taiwan Unix Server market. But because in the bilateral long-term competition situation, in the actual negotiations process, both sides not necessarily are mutually same to all threat cognition, thus creates a threat uncertainty. This research based on the variable-threat bargaining game to analyze both sides how to make a best alternative from the strategic move to reach the bilateral benefit. And according to Thomas Schelling “Brinkmanship”, “the credible commitment”, “the credible deterrence” to propose the suggestions of the bargaining strategy for both sides negotiation. If after both sides (Sun and Fujitsu) reach the agreement, forms the aspect in Taiwan's Unix server market which three stands like a three way dual. This research by the Extensive Games with Perfect Information and sequential game to analyze again three between to compete gathers the relations. Again and compares at the same time the move after to leave move of similarities and differences in turn, proposed tripartite competes best gathers the strategy. But no matter flows out from three way sequential game or simultaneously leaves the move game all to have to consider: (1). The superiority strategy is both two parties (Solaris and HP) is all must leave the move to IBM to obtain the biggest market benefit; (2). The strategy which separates the mountain view tiger is not suiting the three way dual games. The Solaris group is possibly unable to stand by, leaves free IBM and HP mutual competition among, but does not participation. This research hope through the structure of game theory and the inference method, bring new breakthrough and strategic thinking by science for the industry field. Also expect this research can help industry field from zero-sum games competition to non-zero-sum games competition to fight for the extra pay-off.