Answer Happy

Consider a distributed database of video files, where each video file is annotated by keywords (i.e. text). Retrieval is then achieved by using an inverted index that maps keywords to the video files (annotated by those keywords). There is no scoring or ranking and all file matching a search query will be returned. When a single word query is issued, it is looked up independently on thousands of servers each holding a part of the database. Each server returns a list of matching video files. State briefly how this retrieval can be achieved using map and reduce. Explain how your solution can be extended for a multi-word query. Argue that the solution is scalable in both cases. If there are limitation to scalability explain them. [3+4+2=9M]