No abstract

No abstract

DOC (Disconnected Operation Cache) is a two-level client caching system designed and implemented to support connected and disconnected operation of small form-factor, battery powered portable computers. This paper describes some design features of DOC, and our experiences with running it in the software engineering environment of AST Research. The uniqueness of the DOC project stems from the fact that it has been designed to completely support client caching and disconnected operation of currently existing portable notebook computers. The AST's local area network consisting of several Novell servers and top of the line notebooks running MS/DOS 6.2 as clients communicating via an Ethernet connection are used to support the system and perform experiments. DOC supports "connected", "partially disconnected", and "disconnected" modes of operation. The partially disconnected mode of operation has been designed to permit DOC to respond to requests on certain files that are not modified frequently, without server involvement. Results of our experiments indicate not only considerable performance gains for "connected" and "partially disconnected" system but also successful operation in the disconnected mode.

This paper describes design and implementation of an application-aware Informed Prefetching and Caching (IPrC) system for Linux operating system. IPrC is a technique for improving application response time by exploiting I/O and computation parallelism. This proactive mechanism utilizes hints (application disclosed file access patterns) in order to pre-fetch the needed data blocks ahead of time and place them in the page cache. While well-studied in experimental systems IPrC technology has not been transferred to commercial or widely-used operating systems. We believe that our work is unique in that respect. We show that an implementation of the IPrC system in Linux is not only feasible but also extremely beneficial, especially for applications with non-sequential file access patterns. Our IPrC system is implemented by replacing the traditional read-ahead mechanism in the Linux kernel. The experiments conducted on a 60MHz Intel PC show execution time reduction of 15-39% for various testing scenarios.

Data caching in distributed file systems has been studied with regard to performance and availability [7,9,14,15]. The most common place for the cached data is the server's main memory. This type of caching is referred to as server caching, as opposed to client caching when the cached data is stored on the client (remote workstation) site. Although server caching improves performance considerably since it eliminates disk transfer time for each access, it has two drawbacks: it still suffers from a network transfer delay and is very vulnerable to server's failures. Server's failure renders the whole system virtually inoperable.