Thesis (Ph.D.)--Chulalongkorn University, 2011

Since the Internet is a large resource of information that is important to a vast number of people, search engines are implemented for accessing the required resources over the Internet. Every search engine allows users to enter keywords in the search field for searching. Unfortunately, searching by keyword is usually not specific. When users do not know appropriate keywords, the search result is generally a long list of information, much of it irrelevant. Therefore, the protocol named Information Search Protocol (ISP) is proposed. The ISP enables users to enter a partial part of the required URL and keywords; the ISP can narrow down the search. However, the original ISP cannot perform an image search and it has a defect of the routing module that causes network congestion. Therefore, this research proposes the image search mechanism and routing algorithm to apply to the ISP. Since the ISP is modified in all mechanisms, the new version of the ISP is called Specific Information Search Protocol (SISP). The image extraction of the image search mechanism is implemented using Color Region Correlation (CRC). Moreover, this research is also proposing a Patrol Packet (PTP) Algorithm to solve the routing problem. The search result of the SISP is specific. In addition to more effective searches, the PTP can significantly reduce the congestion caused. Therefore, the SISP is a performance search protocol.