Data stores for cloud infrastructures provide limited consistency guarantees, which restricts the applicability of the cloud for many applications with strong consistency requirements, such as financial and medical information systems. Megastore is a replicated data store used in Google’s cloud infrastructure. Data are partitioned into entity groups, and consistency is only guaranteed if each t...

In Multi-writer, Multi-reader systems, data consistency is ensured by the number of replica nodes contacted during read and write operations. Contacting a sufficient number of nodes in order to ensure data consistency comes with a communication cost and a risk to data availability. In this paper, we describe an enhancement of a consistency protocol called LibRe, which ensures consistency by con...

This paper explores ways to provide strong consistency for Internet applications scaling to millions of clients. We make four contributions. First, we identify the ways in which specific characteristics of data-access workloads affect the scalability of cache consistency algorithms. Second, we define two primitive mechanisms, split and join for growing and shrinking hierarchies. We show how the...

Abstract. In this paper we obtain exponential upper bounds for the rate of convergence of a version of the algorithm Context, when the underlying tree is not necessarily bounded. The algorithm Context is a well-known tool to estimate the context tree of a Variable Length Markov Chain. As a consequence of the exponential bounds we obtain a strong consistency result. We generalize in this way sev...

Chain replication is a new approach to coordinating clusters of fail-stop storage servers. The approach is intended for supporting large-scale storage services that exhibit high throughput and availability without sacrificing strong consistency guarantees. Besides outlining the chain replication protocols themselves, simulation experiments explore the performance characteristics of a prototype ...

Consider an i.i.d. sequence of random variables whose distribution f lies in one of a nested family of models Mq , q ≥ 1. We obtain a sharp characterization of the pathwise fluctuations of the generalized likelihood ratio statistic under entropy assumptions on the model classes Mq. Moreover, we develop a technique to obtain local entropy bounds from global entropy computations, so that these re...

We define a solution to the problem of coalition formation that applies to purely hedonic games. Coalition structures satisfying our requirements are called durable, and we interpret them as much more likely to last than those coalition structures not satisfying the requirements, which we call transient. Durability results from a combination of foresight and extreme risk aversion on the part of...

The search for conditions for the consistency of maximum likelihood estimators in nonlinear mixed effects models is difficult due to the fact that, in general, the likelihood can only be expressed as an integral over the random effects. For repeated measurements or clustered data, we focus on asymptotic theory for the maximum likelihood estimator for the case where the cluster sizes go to infin...

An estimator of the Kullback-Leibler information number by using its representation as a functional of the grade density is introduced. Its strong consistency is proved under the mild conditions on the grade density. The same approach is used to study the entropy measure of bivariate dependence (mutual information). Some applications to detection theory are also given.

Abstract: The adaptive control problem for a jump linear system with quadratic cost functional on infinite time interval is solved in this paper. It is assumed that the coefficients of the state equation are unknown but a compact set that contains the parameters is known. A diminishing excitation accompanies the adaptive control signal to ensure the strong consistency of the weighted least squa...