Utility computing achieves efficiencies by dynamically reallocating shared resources between services operating on virtual clusters. These efficiencies can be hard to realize for data intensive applications; newly allocated nodes must be populated with a large amount of data which impedes rapid node reallocation. We describe a data management architecture that uses disk caches on each node to reduce data copying and speed up node reallocation for data intensive applications. Cache consistency management is simplified by extensive use of copy-on-write techniques. A data-driven scheme is then used to select nodes for reallocation between virtual clusters based on the amount of relevant cached data. These nodes are identified using a novel technique of statistically sampling the contents of caches. We demonstrate the benefits of this architecture using our implementation of an efficient block level caching and copy-on-write target for the Linux device-mapper framework

As users acquire or gain access to an increasingly diverse range of web access clients, web applications are adapting their user interfaces to support multiple modalities on multiple client types. User experiences can be enhanced by clients with differing capabilities combining to provide a distributed user interface to applications. Indeed, users will be frustrated if their interaction with applications is limited to one client at a time.This paper discusses the requirements for coordinating web interaction across an aggregation of clients. We present a framework for multi-device browsing that provides both coordinated navigation between web resources and coordinated interaction between variants, or representations, of those resources once instantiated in the clients. The framework protects the application from some of the complexities of client aggregation.We show how a small number of enhancements to the XForms and XML Events vocabularies can facilitate coordination between clients and provide an appropriate level of control to applications. We also describe a novel proxy which consolidates HTTP requests from aggregations of clients and reduces the burden that multi-client browsing places on the application.

There have been considerable advances towards higher speed (100 Mb/s) workgroup LANs which support the existing UTP and STP structured cabling utilized by 10 BASE-T and token ring LANs. The paper describes the transmission techniques used by an IEEE 802.12 demand priority network with UTP and STP structured cabling. The UTP transmission scheme supports categories 3-5 UTP (i.e. voice-grade and data-grade) using a 5B6B block coded binary signalling scheme on four pairs. This binary signalling scheme is shown to provide better immunity against crosstalk and external (impulse) noise than multilevel signalling schemes. The STP scheme combines the strengths of the 5B6B block code with signalling technology similar to existing SDDI links.

Design techniques which enable the performance of subcarrier multiplexed FM broadcast optical networks to be optimized at the planning stage are described. At the transmitter node, criteria for maximizing the subscriber carrier-to-noise ratio (CNR) are presented. At the subscriber node, it is shown that under the majority of circumstances, the optimum direct detection receiver consists of a low-noise III-V avalanche photodiode and high-impedance front-end preamplifier. Sensitivities approaching -40 dB (1 mW) for 16.5 dB CNR in 36 MHz subscriber bandwidth are predicted for a 60-channel system. Preliminary experimental tests on an optical-feedback subcarrier receiver showed -34 dB (1 mW) sensitivity at 1.55 mu m from 16.5 dB CNR in a 36 MHz bandwidth centered on 1.2 GHz with a 12% modulation index single-channel FM test signal.>