Bradley A. Wallace

Simulative performance analysis of gossip failure detection for scalable distributed systems

Three protocols for gossip-based failure detection services in large-scale heterogeneous clusters are analyzed and compared. The basic gossip protocol provides a means by which failures can be detected in large distributed systems in an asynchronous manner without the limits associated with reliable multicasting for group communications. The hierarchical protocol leverages the underlying network topology to achieve faster failure detection. In addition to studying the effectiveness and efficiency of these two agreement protocols, we propose a third protocol that extends the hierarchical approach by piggybacking gossip information on application-generated messages. The protocols are simulated and evaluated with a fault-injection model for scalable distributed systems comprised of clusters of workstations connected by high-performance networks, such as the CPlant system at Sandia National Laboratories. The model supports permanent and transient node and link failures, with rates specified at simulation time, for processors functioning in a fail-silent fashion. Through high-fidelity, CAD-based modeling and simulation, we demonstrate the strengths and weaknesses of each approach in terms of agreement time, number of gossips, and overall scalability.

Surgical therapy for Parkinson's disease.

High frequency stimulation (HFS) has become the main alternative to medical treatment, due to its reversibility, adaptability, and low morbidity. Initiated in the thalamus (Vim) for the control of tremor, HFS has been applied to the Pallidum (GPi), and then to the subthalamic nucleus (STN), suggested by experiments in MPTP monkeys. STN-HFS is highly efficient on tremor, rigidity and bradykinesia and is now widely applied. Criteria for success are correct patient selection and precise electrode placement. The best outcome predictor is the response to Levodopa. The mechanisms of action might associate inhibition of cell firing, jamming of neuronal message and exhaustion of synaptic neurotransmitter release. The inhibition of glutamate STN release could be neuroprotective on nigral cells. Animal experiments support this hypothesis, not contradicted by the long-term follow up of patients. Neuroprotection might have considerable impact on the management of PD patient and warrants clinical trials.

Might deep brain stimulation of the subthalamic nucleus be neuroprotective in patients with Parkinson’s disease?

Abstract Parkinson’s disease (PD) is characterized by nigral degeneration of dopaminergic neurons in the pars compacta of the substantia nigra. Rather than treating only the symptomatic aspects of Parkinson’s disease, one may also consider treatments designed to retard, arrest, or even reverse this degenerative process. Such strategies could include preventive or restorative treatments instead of purely palliative treatments. A recent hypothesis states that glutamate output from the subthalamic nucleus (STN) to the substantia nigra contributes to the neurotoxic process underlying dopaminergic cell death in Parkinson’s disease. Furthermore, high-frequency stimulation (HFS) of the STN inhibits neurons resulting in the suppression of their glutamate output. Experiments in both rats and monkeys provide preliminary data supporting this hypothesis. Kainic acid (KA) lesions of the STN prevent the loss of dopaminergic neurons in the substantia nigra after intrastriatal injection of 6-hydroxydopamine (6-OHDA) in rats, and after systemic administration of MPTP in monkeys. In PD patients, the background level of their disease is evaluated in the off medication/off stimulation state (UPDRS III score), over a period of 5 years. Thirty percent of the patients are stabilized and 18% have persistent improvement of their disease-related impairment. Further experiments are needed, including controlled clinical trials utilizing functional imaging of the dopamine transporters and post-synaptic receptors.

Modeling and Simulative Performance Analysis of SMP and Clustered Computer Architectures

The performance cliaracteris tics of several classes of parallel computing systems are analyzed and com pared using high-fidelity modeling and execution- driven simulation. Processor, bus and network mod els are used to construct and simulate the architec tures of symmetric multiprocessors (SMPs), clusters of uniprocessors, and clusters of SMPs. To demon strate a typical use, the performance of ten systems is evaluated using a parallel matrix-multiplication algorithm. Because the performance of a parallel algo rithm on an architecture depends on its communica tion-to-computation ratio, an analysis of communica tion latencies for bus transactions, cache coherence, and network transactions is used to quantify each systems communication overhead. While low-level performance attributes are difficult to measure on experimental testbed systems, and are difficult to accurately represent in purely analytical models, with high fidelity simulative models they can be readily and accurately obtained. This level of detail allows the designer to rapidly prototype and evaluate the performance of parallel and distributed systems.