E. G. Coman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3(2):67-78, June 1971.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... Why non contracting maps 39 7 Concluding remarks and some open questions 42 1 Introduction Algebraic topological models have been used now for some years in concurrency theory (concurrent database systems and fault tolerant distributed systems as well) The earlier models, progress graph (see [CES71] for instance) have actually appeared in operating systems theory, in particular for describing the problem of deadly embrace 1 in multiprogramming systems . The basic idea is to give a description of what can happen when several processes are modifying shared resources. Given a shared ....

E. G. Coman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3(2):67-78, June 1971.

....in [27] or [14] 2. A topological approach The first algebraic topological model I am aware of is called progress graph and has appeared in operating systems theory, in particular for describing the problem of deadly embrace in multiprogramming systems . Progress graphs are introduced in [9], but attributed there to E. W. Dijkstra. In fact they also appeared slightly earlier (for editorial reasons it seems) in [55] The basic idea is to give a description of what can happen when several processes are modifying shared ressources. Given a shared resource a, we see it as its associated ....

Coffman, E. G., M. J. Elphick and A. Shoshani, System deadlocks, Computing Surveys 3 (1971), pp. 67--78.

.... Why non contracting maps 40 7 Concluding remarks and some open questions 42 1 Introduction Algebraic topological models have been used now for some years in concurrency theory (concurrent database systems and fault tolerant distributed systems as well) The earlier models, progress graph (see [CES71] for instance) have actually appeared in operating systems theory, in particular for describing the problem of deadly embrace 1 in multiprogramming systems . The basic idea is to give a description of what can happen when several processes are modifying shared resources. Given a shared ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3(2):67--78, June 1971.

....If some thread executes a slow (blocking) system call such as sleep or read, then normal process control flow guarantees that only the calling thread is blocked. Bibliographic notes Semaphore operations were proposed by Dijkstra [4] The progress graph concept was introduced by Coffman [3] and later formalized by Carson and Reynolds [2] The book by Butenhof [1] contains a comprehensive description of the Posix threads interface. ....

E. Coffman, M. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3(2):67--78, June 1971.

....is a key issue for the robust and efficient operation of flexibly automated production systems. One of the first realizations in the deadlock related research was the fact that from a strategic conceptual standpoint, deadlocks can be either prevented avoided or detected and recovered [4, 17]. Correct prevention avoidance strategies never allow deadlock to arise by controlling the part flow and the underlying resource allocation according to some deadlock avoidance policy (DAP) that, furthermore, guarantees the liveness of the controlled system [1, 20] 1 On the other hand, the ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

....events. Since the resources involved are returned to the system after they have been used by the acquiring processes, they are characterized as reusable. The abstracted system of interacting processes and the set of shared resources will be characterized as a Resource Allocation System (RAS) In [15, 5] the necessary and sufficient conditions for the development of a deadlock in a RAS are identified. These conditions suggest possible methodologies for addressing the problem of deadlock. These methodologies can be broadly classified as: i) Prevention, ii) Detection Recovery, and (iii) ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

....execution. 1 The study of resource allocation models was initiated in the field of computer system engineering in the late 60s and early 70s, in the context of a broader effort in providing robust and efficient operating systems [6, 9, 10, 11] It was quickly realized and formally established [5, 11] that most of the relevant resource allocation environments give rise to situations in which a number of processes get into a circular wait state where they wait for each other to release resources required for the execution of their next processing stages; this situation was defined as a system ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

....access to a vessel currently used by the other recipe. This might lead to a loss of an entire batch, and must be avoided. Such a situation is called a circular wait, in the following we will refer to this as a deadlock situation. In general there are four necessary requirements for a deadlock, see [2], but in batch processing systems only circular waits are of interest, because the other three requirements normally always hold. This paper describes the implementation of deadlock avoidance capabilities to SattBatch. The methodology is general and can be applied to other control systems as ....

E.G. Coffman, M.J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3(2):67--78, 1971.

....discussed herein to higher level models in [55, 48] 2.3 The nature of the RAS deadlock and generic resolution approaches The problem of system deadlock was first studied in the context of Computer System Engineering, back in the late 60s and early 70s. The work presented in papers like [27, 26, 10, 28] set the foundation for understanding the problem nature. Specifically, one of the early findings in the study of reusable RAS [10] was that for the development of a deadlock in these systems, the simultaneous satisfaction of the next four conditions is necessary and sufficient: mutual exclusion ....

....of system deadlock was first studied in the context of Computer System Engineering, back in the late 60s and early 70s. The work presented in papers like [27, 26, 10, 28] set the foundation for understanding the problem nature. Specifically, one of the early findings in the study of reusable RAS [10] was that for the development of a deadlock in these systems, the simultaneous satisfaction of the next four conditions is necessary and sufficient: mutual exclusion i.e. tasks must claim exclusive control of the resources that they require, wait for condition i.e. tasks hold resources ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

....and deadlock related properties. Resource allocation systems and the underlying problem of deadlock avoidance arising from conflicting resource requests, were initially studied in the context of computer operating systems, back in the late 60 s and early 70 s; the work presented in, e.g. [15, 8, 16], is indicative of that effort. Subsequently, the problem complexity was formally characterized in late 70 s [2, 14] while in the 80 s, deadlock resolution was addressed in the context of distributed databases and computation, e.g. 34] Since early 90 s there is a resurgent interest in the ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

....events. Since the resources involved are returned to the system after they have been used by the acquiring processes, they are characterized as reusable. The abstracted system of interacting processes and the set of shared resources will be characterized as a Resource Allocation System (RAS) In [15, 5] the necessary and sufficient conditions for the development of a deadlock in a RAS are identified. These conditions suggest possible methodologies for addressing the problem of deadlock. These methodologies can be broadly classified as: i) Prevention, ii) Detection Recovery, and (iii) ....

E. G. Coffman, M. J. Elphick, and A. Shoshani. System deadlocks. Computing Surveys, 3:67--78, 1971.

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