Mancini, L., Pappalardo G.: Towards a theory of replicated processing. Lecture Notes in Computer Science 331, Springer-Verlag, 175-192 (1988).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....replication, such as primarybackup [BMST93] and active replication [Sch93] However, these are specifications of replication solutions rather than a specification of the actual problem solved by replication. The very few abstract replication properties that we know about, e.g. Aiz89] and [MP88], do not address correctness with respect to external side effect. They only address consistency of state that is encapsulated within the service itself. In particular, there is no provisioning, in the specifications, for having a replicated service call a third party entity, e.g. another ....

....Considering that a replicated program is correct if it can somehow be shown to be equivalent to a non replicated program is an intuitive idea, and this idea has already been explored by different authors. The definition of equivalence is the main difference between the various approaches: ffl In [MP88], an algebra of action sequences is used to define a correctness criterion for replication. The N replication of a base process is a replicated process, denoted by P N . The replicated process P N is correct if it is possible to extract , from every trace t N of P N , a trace t of P . The ....

[Article contains additional citation context not shown here]

L. Mancini and G. Pappalardo. Towards a theory of replicated processings. In Formal Techniques in Real-time and Faulttolerant Systems, pages 175--192. LNCS (331), Springer Verlag, 1988.

....replication, such as primary backup [BMST93] and active replication [Sch93] However, these are specifications of replication schemes rather than specification of the actual problem solved by replication. The very few abstract replication properties that we know about, e.g. Aiz89] and [MP88] do not address correctness with respect to external side e#ect. They only address consistency of state that is encapsulated within the service. In particular, there is no provisioning, in the specifications, of having a replicated service call a third party entity, e.g. another replicated ....

....replicated program is correct if it can somehow be shown to be equivalent to a non replicated program is an intuitive idea, and this idea has already been explored by di#erent authors. The main di#erences between the various approaches have to do with how the notion of equivalence is defined. In [MP88] an algebra of action sequences is used to define a correctness criteria for replication. The N replication of a base process is a replicated process, denoted by P N . The replicated process P N is correct if it is possible to extract, from every trace t N of P N , a trace t of P . The ....

[Article contains additional citation context not shown here]

L. Mancini and G. Pappalardo. Towards a theory of replicated processings. In Formal Techniques in Real-time and Fault-tolerant Systems, pages 175--192. LNCS (331), Springer Verlag, 1988.

.... was never completely verified [17] the concept of Byzantine Generals algorithms was developed [18] as was the first fault tolerant clock synchronization algorithm with a mathematical performance proof [9] Other theoretical investigations have also addressed the problems of replicated systems [19]. Some recent work at SRI International has focused on problems related to the style of fault tolerant computing adopted by RCP. Rushby has studied a fault masking and transient recovery model and created a formalization of it using Ehdm [20, 21] In addition, Shankar has undertaken the ....

Mancini, L. V.; and Pappalardo, G.: Towards a Theory of Replicated Processing. In Lecture Notes in Computer Science, vol. 331, pp. 175--192. Springer Verlag, 1988.

.... was never completely verified [10] the concept of Byzantine Generals algorithms was developed [7] as was the first fault tolerant clock synchronization algorithm with a mathematical performance proof [6] Other theoretical investigations have also addressed the problems of replicated systems [8]. 4 Application Definition We present a method for specifying an operating system workload that characterizes the interface between the application software and the operating system. The specification consists of a generic set of mathematical definitions serving as a schema. For an actual ....

Luigi V. Mancini and Giuseppe Pappalardo. Towards a theory of replicated processing. In Lecture Notes in Computer Science, volume 331. Springer Verlag, 1988.

.... [NASA 1983] the concept of Byzantine Generals algorithms was developed [Lamport 1982] as was the first fault tolerant clock synchronization algorithm with a mathematical performance proof [Lamport 1985] Other theoretical investigations have also addressed the problems of replicated systems [Mancini 1988]. Some recent work has focused on problems related to the style of fault tolerant computing adopted by RCP. Rushby has studied a fault masking and transient recovery model and created a formalization of it using Ehdm [Rushby 1991, Rushby 1992] Rushby s model is more general than ours, but assumes ....

L. V. Mancini and G. Pappalardo. Towards a theory of replicated processing. In Lecture Notes in Computer Science, volume 331, pages 175--192. Springer Verlag, 1988.

....and Traces, where the second and third terms come from the CSP model [1, 2, 3] on which it is based, and the first refers to a specific technique used to relate systems specified at different levels of abstraction. The ERT model originates from previous work on distributed replicated systems [5, 10], which has led to the development of a formal model [6, 7, 8] based on an implementation relation capturing the notion that a replicated system is a correct implementation of another target or base system. The relation was defined independently of replication, in an abstract way, which enabled a ....

Mancini, L., Pappalardo G.: Towards a theory of replicated processing. Lecture Notes in Computer Science 331, Springer-Verlag, 175-192 (1988).

....clock implements a single, precise clock To our knowledge the proposed problem has been paid scant attention so far in the literature. Yet, it has a clear practical interest, e.g. extending to some fundamental fault tolerance techniques. Indeed, restricted versions of the treatment proposed [MP88,KMP97] have been applied to the analysis of NMR (N Modular Redundancy) a fault tolerance technique whereby N possibly faulty copies of a base system are assembled and coordinated so as to mimic the behaviour of a single non faulty copy. Another interesting application [KP98] has been the ....

....In contrast, note that our treatment will not even require any combining device except process composition and renaming. The FD model adopted is expressive, well understood and provides a wealth of important theoretical results. A treatment could also be given within a pure trace model [MP88] but this would only allow partial (functional) correctness to be studied, whereas failures afford the ability to describe internal nondeterminism (hence deadlock properties) Divergences are traces after which a process might become uncontrollable, or diverge. While their presence can be ....

[Article contains additional citation context not shown here]

L. V. Mancini and G. Pappalardo. Towards a theory of replicated processing. In M. Joseph, editor, Formal Techniques in Real-Time and Fault-Tolerant Systems, Lecture Notes in Computer Science, vol. 331, pages 175--192, Warwick, 1988. Springer-Verlag.

No context found.

L. V. Mancini and G. Pappalardo. Towards a Theory of Replicated Processing. In M. Joseph, editor, Proceedings of the symposium on Formal Techniques in Real-Time and Fault-Tolerant Systems: LNCS 331, pages 175#192. Springer Verlag, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC