各種物聯網應用的出現加速了「語意互操作性」的引入。為了因應物聯網所生成的大量數據,「語意互操作性」嘗試使用語意資訊(即本體,或稱本體論、知識本體)註釋這些資料,而不存儲龐大的原始數據,藉以實現資料的智慧管理。「語意推論」是本體的其中一個主要功能,得以從明確的上下文資訊中,推導出隱含的知識。然而,在雲端的架構下,語意推論在處理巨大的資料量時,會耗用大量的計算資源。相對於集中式的雲端架構,在本研究中,我們提出「霧運算」的採用,於分散式霧節點的階層式架構中,分布整個語意推論的程序。為了驗證方法的有效性,我們以老人照護作為使用案例,將分散式的霧架構與集中式的雲端架構進行比較,測試兩者的整體效能及優劣。
The emergence of various Internet of Things (IoT) applications has hastened the introduction of Semantics Interoperability (SI). Instead of storing enormous raw data, SI attempts to annotate these data with semantic information (a.k.a. ontology) that can enable smart management of the ever-increasing data generated by IoT. One of the main features of ontology is semantic reasoning for deriving implicit knowledge from explicit context information. Nevertheless, such semantic reasoning in the cloud can be computationally expensive with the gigantic amount of incoming data. In this research, we propose the adoption of "Fog Computing" to distribute the procedure of semantic reasoning among a hierarchy of distributed Fog nodes than in a centralized cloud. To verify the effectiveness of our methods, we compare the proposed distributed Fog architecture with a centralized Cloud system based on the use case of elderly care.