Comment: They characterize four parallel applications: sort, matrix multiply, seismic migration, and video server, in terms of their I/O activity.

Abstract: With a vast number of resources being committed to reformatting into digital form, we need to consider how we can ensure that digital information will continue to be accessible over a prolonged period of time. This chapter first outlines the general problem of information in digital form disappearing. It then looks closely at five key factors that pose problems for digital longevity. Finally, we turn our attention to a series of suggestions that are likely to improve the longevity of digital information, focusing primarily on metadata. This chapter was written for the library, museum, and archives communities. However, the observations will be useful for all communities wishing to ensure the longevity of any type of digital information.

Abstract: The promise has been made: ``Digital information is forever. It doesn't deteriorate and requires little in the way of material media.'' So said one of the chieftains of the emerging digital age, computer-chip maker Andy Grove, the head of the Intel Corporation. Another chieftain, Librarian of Congress James H. Billington, has set about digitizing the world?s largest library so that its contents can become accessible by anyone, from anywhere, forever. But a shadow has fallen. ``It is only slightly facetious,'' wrote RAND researcher Jeff Rothenberg in Scientific American, ``to say that digital information lasts forever—or five years, whichever comes first.''

Abstract: Managing large objects with high data-rate requirements is difficult for current computing systems. We describe an Input/Output architecture, called Swift, that addresses the problem of storing and retrieving very large data objects from slow secondary storage at very high data-rates. Applications that require this capability are poorly supported in current systems, even though they are made possible by high-speed networks. These range from storage and visualization of scientific computations to recodring and play-back of color video in real-time. Swift addresses the problem of providing the data rates required by digital video by exploiting the available interconnection capacity and by using several slower storage devices in parallel. We have done two studies to validate the Swift architecture: a simulation study and an Ethernet-based proof-of-concept implementation. Both studies indicate that the aggregation principle proposed in Swift can yield very high data-rates. We present a brief summary of these studies.