Matteo Sereno

Fine grained software degradation models for optimal rejuvenation policies

In this paper, we address the problem of preventive maintenance of operational software systems, an approach recently proposed to counteract the phenomenon of software “aging”. We consider the so-called “software rejuvenation” technique [Software rejuvenation: analysis, module and applications, in: Proceedings of the 25th International Symposium on Fault-Tolerance Computing (FTCS-25), Pasadena, CA, USA, June 1995], which consists in periodically stopping the software system and then restarting it in a “robust” state after a proper maintenance, and we propose a methodology for the quantitative analysis of rejuvenation strategies. The methodology is based on a fine grained model that assumes that it is possible to identify the current degradation level of the system by monitoring an observable quantity, so that the rejuvenation strategy can be tuned on the measured degradation. Based on this methodology, we present two different strategies that allow to decide whether and when to rejuvenate, and we exploit the theory of renewal processes with reward to estimate the steady-state unavailability of the software system, which is used to define an optimality criterion that allows us to evaluate the proper rejuvenation intervals. The methodology and the rejuvenation strategies are demonstrated by applying them to a real-world case study, arising in the area of database maintenance for data archiving, and to a hypothetical setting used to assess the sensitivity of the technique to various degradation processes.

Fluid Stochastic Petri Nets Augmented with Flush-out Arcs: Modelling and Analysis

Fluid(or Hybrid) Petri Nets are Petri net based models with two classesof places: discrete places that carry a natural number of distinctobjects (tokens), and fluid places that hold a positive amountof fluid, represented by a real number. With respect to previousformulations, the FSPN model presented in this paper, is augmentedwith a new primitive, called flush-out arc. A flush-out arc connectsa fluid place to a timed transition, and has the effect of instantaneouslyemptying the fluid place when the transition fires. The paperdiscusses the modeling power of the augmented formalism, andshows how the dynamics of the underlying stochastic process canbe analytically described by a set of integro-differential equations.A procedure is presented to automatically derive the solutionequations from the model specifications. The whole methodologyis illustrated by means of various examples.

On the fly gaussian elimination for LT codes

We propose an improved algorithm for decoding LT codes using Gaussian elimination. Our algorithm performs useful processing at each coded packet arrival thus distributing the decoding work during all packets reception, obtaining a shorter actual decoding time. Furthermore, using a swap heuristic the decoding matrix is kept sparse, decreasing the cost of both triangularization and back-substitution steps.

Generalized Probabilistic Flooding in Unstructured Peer-to-Peer Networks

In this paper, we propose a generalization of the basic flooding search strategy for decentralized unstructured peer-to-peer (P2P) networks. In our algorithm a peer forwards a query to one of its neighbors using a probability that is a function of the number of connections in the overlay network of both. Moreover, this probability may also depend on the distance from the query originator. To analyze the performance of the proposed search strategy in heterogeneous decentralized unstructured P2P networks, we develop a generalized random graph (GRG)-based model that takes into account the high variability in the number of application level connections that each peer establishes, and the nonuniform distribution of resources among peers. Furthermore, the model includes an analysis of peer availability, i.e., the capability of relaying queries of other peers, as a function of the query generation rate of each peer. Validation of the proposed model is carried out comparing the model predictions with simulations conducted on real overlay topologies obtained from crawling the popular file sharing application Gnutella.

On the Product Form Solution for Stochastic Petri Nets

The combinatorial explosion of the state space of Stochastic Petri Nets (SPNs) is a well known problem that inhibits the exact solution of large SPNs, and therefore a broad use of this kind of Petri Nets as a modelling tool. The same problem exists also for other modelling formalisms like for example Queueing Networks (QNs). In [13, 3] a class of QNs whose solution can be computed in an easy way was defined. For this class of models the solution can be factorized into terms that refer to each single queue of the network. This solution is known as Product Form Solution (PFS).

Analysis of PPLive through active and passive measurements

The P2P-IPTV is an emerging class of Internet applications that is becoming very popular. The growing popularity of these rather bandwidth demanding multimedia streaming applications has the potential to flood the Internet with a huge amount of traffic.

Fluid models for large-scale wireless sensor networks

We develop an analytical framework for the study of large-scale, wireless sensor networks. We use a fluid approach, i.e., we represent the sensor network by a continuous fluid entity distributed on the network area. We assume that sensors employ a CSMA/CA mechanism to access the channel and a minimum energy consumption criterion to select traffic routes, and accurately model these aspects. Furthermore, the framework accounts for energy consumption and is able to model the active/sleep dynamics of the nodes. We validate our approach through simulation and show that the proposed framework is well suited for describing the properties of large sensor networks and understanding their complex behavior. To the best of our knowledge, our framework is one of few analytical models developed for large-scale sensor networks which can accurately capture the main aspects of these systems.

Modeling software systems with rejuvenation, restoration and checkpointing through fluid stochastic Petri nets

In this paper, we present a Fluid Stochastic Petri Net (FSPN) based model which captures the behavior of aging software systems with checkpointing, rejuvenation and self-restoration, three well known techniques of software fault tolerance. The proposed FSPN based modeling framework is novel in many aspects. First, the FSPN formalism itself is extended by adding flush-out arcs. Second, the three techniques are simultaneously captured in a single model for the first time. Third, the formalism enables modeling dependencies of the three techniques on various system features such as failure, load and time in the same framework. Further, our base FSPN model can be viewed as a generalization of most previous models in the literature. We show that these FSPNs can not only mimic previously published models but can also extend them. For one FSPN model, we present numerical results to illustrate their usage in deriving measures of interest.

A simple analytical framework to analyze search strategies in large-scale peer-to-peer networks

This paper presents an analytical framework to study search strategies in large-scale decentralized unstructured peer-to-peer (P2P) networks. The peers comprising the P2P network and their application-level connections are modeled as generalized random graphs (GRGs) whose simple and efficient analysis is accomplished using the generating function of the graphs degree distribution. The framework we defined allows the computation of several interesting performance indexes to be used to compare different search strategies: in particular, the average number of messages sent throughout the P2P network and the probability that a query is successful are used as examples. Furthermore, assuming that the cumulative distribution function (CDF) of the time required by a peer to positively reply to a query is known, we show how to derive the CDF of the time it takes for a randomly chosen peer to obtain at least one positive reply from other peers. The approach is validated through simulation showing that the accuracy of the proposed model improves as the size of the P2P network increases making it a suitable tool for the analysis of search strategies in large-scale systems.

Fluid stochastic Petri nets: An extended formalism to include non-Markovian models

A Fluid Stochastic Petri Net (FSPN) formalism, where there are two kind of places, one which carries discrete tokens and the other which contains continuous quantity, is presented and discussed. In the proposed formulation, a new primitive is introduced, called flush-out arc. A flush-out arc connects a transition to a continuous place and has the effect of instantaneously empty the place when the transition fires. With this extension, FSPNs can be viewed as a graphical formalism to represent stochastic models with reward rates that can be dependent on the discrete as well as the continuous component of the state space descriptor. First the model is formally introduced and the integro-differential equations representing the dynamic of the system are fully derived in the case of a single continuous place. However, the goal of the paper is to propose a first step towards the automatic solution of a general FSPN model starting from its graphical description. Furthermore, in order to illustrate the potentiality of the approach, we show that the proposed formalism is suited to convert a non-Markovian SPN, of the type considered up to now in the literature, into a FSPN. Since, however, the FSPN formalism is more general and flexible, various modeling extensions, not conceivable in the non-Markovian SPN setting, are investigated.