Revision with unchanged content. A data stream being transmitted over a network channel with capacity less than the data transmission rate of the data stream causes sequential network problems. In this book, we present a new approach for shedding less-informative attribute data from a data stream to maintain the data transmission rate less than the network channel capacity. A scheme for shedding attributes, instead of tuples, becomes imperative in stream data, since shedding a complete tuple would lead to shedding some informative, as well as less-informative, attribute data in the tuple. Since data shed at the source site may be of interest to the user at the destination site, we design a data recovery approach, which maintains the minimal amount of information for data recovery purpose while imposing minimal overhead for data recovery on the source site. Our load shedding and data recovery approach (i) handles wide range of data streams in different application domains, (ii) is dynamic in nature, since each load shedding scheme adjusts the amount of data to be shed according to the current load and network capacity, and (iii) is adoptive, which is appealing in an ever-changing network environment, and (iv) is not based on queries, but works on general data streams instead.