This presentation briefly reviews the knowledge management (KM) topic, and prodives an example of how issues around KM is effectively solved in a Fortune 500 company. Knowledge has a very strong human element [9,10]. Hence, researchers have identified two classses of knowledge: soft (tacit) and hard (explicit) [5,7]. Hard knowledge can be codified, and documented. Soft knowledge cannot be externalized. As a result, effective KM in large, geographically dispersed companies, should not only include central repository of documents, but also managing the human element as well [3]. To address the human element, researchers defined "communities of practice (CoP)" where co-located individuals share their knowledge typically through face-to-face communication [6,13]. Similarly "network of practice (NoP)" concept 12,111 expanded the CoP to cover knowledge sharing among individuals who are not co-located. The relationships in NoP are not as strong as that in CoP. Further, communication in NoP is typically not face-to-face, and could take place in electronic medium. Another aspect of knowledge sharing in a loosely coupled network is to realize the network effects. Reed's 181 and Metcalf's 141 laws state that the value of a network increases at a much higher rate than the number of users increase: exponential and quadratic, respectively. There is a recent study [1] that claims that the value of a network increases less gradually. To realize the network effects, though, the communication medium has to be peer-to-peer and not centralized [12]. Because with centralized communication model, benefits of network effect diminishes significantly. Having provided a high level overview of KM, we defined the problem in EMC Corporation in late 1990s. A small group of peformance engineers were providing expertise to a small group of field and sales engineers. As more field and sales engineers are interested in and requested performance expertise, it became evident that this model would not scale. As a result, the company started a new vitual community (or network) and invited all field and sales engineers who are interested in performance related topics. This new network, called SPEED, would provide communication among individuals with similar interests, and it would simplify dissemination of knowledge and experience of the performance engineering group in corporate headquarters. SPEED addressed both hard and soft elements of knowledge. All performance related data are published in a central database where all members can access. Email exchanges among members are stored in webboard format so that members can search on previous topics. Performance engineering holds annual conferences to update field and sales engineers with up-to-date information. Between conferences, they provide Internet-based seminars and welicasts to provide latest developments. Even though all emails are stored centrally, communication among members are never centralized. Peer-topeer communication is always available, and in fact is encouraged. Members have a very strong motivation to join SPEED virtual community. SPEED membership is an important recognition for field and sales engineers as they have to be nominated by their management, and have to prove their competence by passing a certification (entrance) exam. Escalation proceess of SPEED makes it easy to scale out to serve thousands of field and service engineers. Currently, about 25 performance engineers support more than 800 SPEED gurus, and these gurus support thousands of field and sales engineers in all geographies of the company. Within the company, SPEED is considered a big success story since it leverages expertise of a small number of engineers to a much bigger audience. It further provides a virtous cycle in that members provide feedback to performance engineering with their hands-on field experience. Looking at SPEED through KM perspective, we can identify reasons for its huge success. First it addressed both the soft and hard aspects of knowledge. Second, it created a Network of Practice, where people with similar backgrounds and interests shared their experiences. Further, SPEED leveraged the network effect. As more members participated, the value of the network for everyone has increased. Lastly, the communication model was not centralized. Members were encouraged to communicate among themselves. SPEED was originally established for the only product line of the company. Since then, the company released or acquired more product lines. Based on the success of the model, the company has established parallel SPEED structures for other product lines. It is hoped that the knowledge managemend model of SPEED would prove to be successful with these product lines as well.
