We describe the support to object management of the distributed object-oriented operating system SOS. We discuss the integration of object migration and storage into C++ programs, a language not designed for that purpose. The necessary support is split between the compiler and a run-time object management system. Migration and storage preserve the type and structure of the objects, which may be user-defined and arbitrarily complex. Our mechanisms are simple, generic, and require little programmer intervention. The key elements are dynamic classes, a generalized pointer type which allows to efficiently refer to an object, pre-requisite objects, and a mechanism for object re-initialization. Keywords: persistence, migration, dynamic linking, dynamic type-checking, C++, object-oriented operating system, SOS 1 INTRODUCTION We present some aspects of the object management support in the SOS operating system [Sha86a]. SOS is a research program to build a distributed operating system, where ...

Based on the notion of persistent threads in Tycoon [MS94], we investigate thread migration as a programming construct for building activity-oriented distributed applications. We first show how a straight-forward extension of a higher-order persistent language can be used to define activities that span multiple (semi-) autonomous nodes in heterogeneous networks. In particular, we discuss the intricate binding issues that arise in systems where threads are first-class language citizens that may access local and remote, mobile and immobile resources. We also describe how our system model can be understood as a promising generalization of the more static architecture of first-order and higher-order distributed object systems. Finally, we give some insight into the implementation of persistent and migrating threads and we explain how to represent bindings to ubiquitous resources present at each node visited by a migrating thread on the network to avoid excessive communication or storag...

. Larchant-RDOSS is a distributed shared memory that persists on reliable storage across process lifetimes. Memory management is automatic: caching of data and of locks, coherence, collecting objects unreachable from the persistent root, writing reachable objects to disk, and reducing store fragmentation. Memory management is based on a novel garbage collection algorithm, that (i) approximates a global trace by a series of partial traces within dynamically determined subsets of the memory, (ii) causes no extra I/O or locking traffic, and (iii) needs no extra synchronization between the collector and the application processes. This results in a simple programming model, and expected minimal added application latency. The algorithm is designed for the most unfavorable environment (uncontrolled programming language, reference by pointers, non-coherent shared memory) and should work well also in more favorable settings. 1 Introduction The Reliable Distributed Object Storage System (Larcha...

We study tracing garbage collection (GC) for a distributed shared memory (DSM) in order to provide persistence by reachability (PBR), in a largescale distributed system. Within a general model of DSM, we specify a distributed tracing GC algorithm that scales, collects cycles, and is orthogonal to coherence. Its main features are: (i) piecewise collection of opportunistically-chosen subsets of the memory, (ii) each site collects independently of other sites, (iii) data replicas are collected independently and no coherence operation is needed for GC purposes; and (iv) asynchrony of collection with respect to applications. 1 Introduction A basic function of an operating system is the sharing of information among its application processes. Larchant supports sharing via a persistent virtual memory, shared by applications even if they run at different sites and/or at different times. From the point of view of the application programmer, persistence, memory management, and distribution are ...

This paper presents the solutions Larchant proposes to these problems. The Larchant GC is a novel hybrid of tracing and counting. It traces whenever economically feasible, i.e., as long as the trace remains local to a site, and reference-counts references that would cost I/O or network traffic to trace.

A number of actions, collectively known as binding, prepare a reference for invocation of its target: locating the target, setting up a connection, checking access rights and concurrency control state, type-checking, instantiating a proxy, etc. Existing languages or operating systems support only a single binding policy, that cannot be tailored to object-specific semantics for the management of distribution, replication, or persistence. We propose a general binding protocol covering the above needs; the protocol is simple (a single RPC and one upcall at each end) but recursive; however the recursion can be terminated at any point, trading off simplicity and performance against completeness. This comprehensive, unified protocol is capable of supporting different languages and object models, and may be tailored to support various policies in a simple manner. 1 Introduction Any computer system supports some reference mechanism (such as ports, sockets, file descriptors, UIDs, capabilities...

We call the older hierarchical and network systems first generation database systems and refer to the current collection of relational systems as the second generation. In this paper we consider the characteristics that must be satisfied by the next generation of data managers, which we call third generation database systems. Our requirements are collected into three basis tenets along with 13 more detailed propositions. 1.