International Computer Science Institute, Berkeley, CA 2007 – Present Postdoctoral Researcher, Networking Group, working with Prof. Scott Shenker, Dr. Petros Maniatis, and Dr. Sylvia Ratnasamy Diverse replication for single-machine Byzantine-fault tolerance: Investigate exploiting cores in many-core systems to defend against software attacks. Explore different isolation and software diversity m...

The ability to migrate hardware processes in a network of hardware reconfigurable nodes improves the fault tolerance of these networks. The degree of fault tolerance is inherent to such networked systems and can be optimized during design time. Therefore, an efficient way of calculating the degree of fault tolerance is needed. This paper presents an approach based on satisfiability testing whic...

Fault tolerance is the ability of a system to continue delivering its services after faults have caused errors. We have argued, in the past, that complex and/or critical systems are best validated by a wide range of methods, including proving, testing, and fault tolerance; we have also argued that in order to use these methods in concert, we need to cast them in a common framework. In this pape...

Discrete controller synthesis (DCS) is a formal approach, based on the same state-space exploration algorithms as model-checking. Its interest lies in the ability to obtain automatically systems satisfying by construction formal properties specified a priori. In this paper, our aim is to demonstrate the feasibility of this approach for fault tolerance. We start with a fault intolerant program, ...

This paper investigates fault tolerance in feedforward neural networks, for a realistic fault model based on analog hardware. In our previous work with synaptic weight noise we showed significant fault tolerance enhancement over standard training algorithms. We proposed that when introduced into training, weight noise distributes the network computation more evenly across the weights and thus e...

The designation “fault tolerant software” has been used for techniques ranging from roll-back and retry to N-version programming, from data mirroring to functional redundancy. If the term is to be meaningful, qualifying definitions are required. This paper attempts to provide these by analyzing the capabilities of representative software fault tolerance techniques described in prior literature ...

Fault tolerance must be included in the initial design of real time systems, must encompass hardware and software, and must be integrated with timing constraints. In many situations, the fault-tolerant design must be static, due to extremely high data rates and severe timing constraints. Ultrareliable systems need to employ proof-of-correctness techniques to ensure fault tolerance properties. W...

Replication is widely used to improve fault tolerance in distributed and multi-agent systems. In this paper, we present a different point of view on replication in multi-agent systems. The approach we propose is based on role concept. We define a specific “fault tolerant role” which encapsulates all behaviors related to replication-based fault tolerance in this work. Our strategy is mainly focu...

A middleware architecture named ROAFTS (Real-time Object-oriented Adaptive Fault Tolerance Support) is presented. ROAFTS is designed to support adaptive fault-tolerant execution of not only conventional process-structured distributed RT application software but also new-style object-structured distributed RT application software. While ROAFTS contains fault tolerance schemes devised for quantit...