Efficient RMI: dynamic specialization of object serialization

Abstract

This paper describes a novel approach to object serialization in remote method invocation (RMI). Object serialization transforms objects' representations between heterogeneous platforms. Efficient serialization is primary concern in RMI because the conventional approaches incur large runtime overheads. The approach described here specializes a serializing routine dynamically according to a receiver's platform, and this routine converts the sender's in-memory representations of objects directly into the receiver's in-memory representations. This approach simplifies the process of RMI: the receiver can access the passed objects immediately without any data copies and data conversions. A new platform can join the existing community of senders and receivers because a specialized routine for the platform is generated as needed. Experimental results show that significant performance gains are obtained by this approach. The prototype implementation of this approach was 1.9 -3.0 times faster than Sun XDR, and the time needed for generating a specialized routine was only 0.6 msec.

Enhancing Remote Method Invocation through Type-Based Static Analysis

“…Braux [1] exploits static analysis to reduce the computational overhead of an invocation due to the reflective calls needed to discover the dynamic type information. The work of Kono and Masuda [7] relies on the existence of run-time knowledge about the receiver's platform, and redefines the serialization routine accordingly. On the sender side, the object to be serialized is converted directly into the receiver's in-memory representation, so that the receiver can access it immediately without any data copy and conversion.…”

“…An overhead is introduced in terms of both computation, as (de)serialization requires a recursive navigation of the object graph, and communication, as large objects result in large serialization streams being transmitted. Most of the existing approaches focus on reducing only the computational overhead [8,1,7,2,12]. They aim at improving the middleware run-time without considering the application code exploiting it and, in particular, how the object is used after deserialization.…”

An Approach for Distributed Objects Applied to Satellite Simulator

“…-Interoperability: This important characteristic allows the distributed object other methods, even if platforms, operational system and programming language are heterogeneous, which can resolve the restrictions imposed by the existing heterogeneity in the Satellite Simulator. The interoperability a significant property that distributed structures must assure, since not all the computer set need to use heterogeneous platforms [7].…”

Towards bridging the gap between programming languages and partial evaluation

“…Numerous strategies [5,9,12,15] have been developed for the specialization process, leading to a wide variety of effects that can be achieved by partial evaluation; many of these techniques have been implemented in practical partial evaluators. These tools have been applied to a wide range of realistic applications in domains such as operating systems [17,23], scientific computing [4,18], graphics [2,16] and software engineering [21,28].…”