Showing papers by "Hans P. Reiser published in 2005"

A flexible and extensible object middleware: CORBA and beyond

Abstract: This paper presents a CORBA-compliant middleware architecture that is more flexible and extensible compared to standard CORBA The portable design of this architecture is easily integrated in any standard CORBA middleware; for this purpose, mainly the handling of object references (IORs) has to be changed To encapsulate those changes, we introduce the concept of a generic reference manager with portable profile managers Profile managers are pluggable and in extreme can be downloaded on demand To illustrate the use of this approach, we present a profile manager implementation for fragmented objects and another one for bridging CORBA to the Jini world The first profile manager supports truly distributed objects, which allow seamless integration of partitioning, scalability, fault tolerance, end-to-end quality of service, and many more implementation aspects into a distributed object without losing distribution and location transparency The second profile manager illustrates how our architecture enables fully transparent access from CORBA applications to services on non-CORBA platforms

A flexible replication framework for scalable and reliable .net services.

Abstract: We present a flexible framework for the development of reliable and scalable distributed services based on the Microsoft .NET framework. Scalability is obtained by load balancing among dynamically managed service replicas. Fault tolerance is achieved by redundancy among replicas and an active replication strategy to ensure replica consistency. The framework architecture features full replication transparency for client applications. True object-oriented programming is supported in the sense that remote references to replicated services may be passed transparently as serialised arguments in remote invocations. We compare our framework to other replication architectures and illustrate the performance with run-time measurements.

A Reconfigurable System Architecture for Consensus-based Group Communication.

Abstract: Using distributed consensus algorithms is an established way to implement group communication systems with total message ordering. This paper presents a system architecture focusing on (re-)configurability of such systems. Unlike other systems, ours can be configured to work with different failure models and low-level communication protocols without changing the application. Failure models, protocols, and their run-time parameters (e.g., timeouts) can even be re-configured at run-time without loosing consistency, especially in case of failures. Re-configuration at run-time allows for applications that adapt to access patterns and environment conditions for gaining optimal performance and fault tolerance at the same time. As reconfiguration is transparent to the application logic, it may be done automatically by the system. We also present performance figures of our prototype implementation on the basis of variants of the Paxos algorithm.

Stable, time-bound object references in context of dynamically changing environments

Abstract: Most middleware platforms lack sufficient support to provide reliable references for accessing distributed mobile objects in the context of dynamic environments, or they heavily depend on infrastructure mechanisms like a globally available location service. Based on the fragmented-object model and on our own middleware implementation AspectIX, this paper presents concepts and mechanisms that avoid outdated references for dynamically distributed objects without additional infrastructure services. Our approach is based on time-bound guarantees. Even in the context of malicious nodes and a partial invalidation of references a safe binding of the distributed object is guaranteed if ever possible.