M. Singhal, N Shivara t ri, Advanced Concepts in Operating Systems, Mc Graw Hill 1994. p.98.   Home/Search   Document Not in Database   Summary   Related Articles   Check

....usage of the large computing capacity of a distributed system depends on the success of its resource management system. A distributed process scheduler manages the resources of the whole system efficiently by distributing the load among the processors to maximize the overall system performance[1]. The distributed scheduler must perform the load distributing operations transparently, which means the whole system is viewed as a single computer by the users of it [2] A distributed system consists of independent workstations connected usually by a local area network. Users of the system ....

Singhal M., Shivaratri N., 1994, Advanced Concepts In Operating Systems, McGraw Hill

....of Permissions. A Permission is a name and a list of allowed actions. For example, a file permission consists of a pathname and a list of actions like read and write . Normal implementation of sandboxing results in the very familiar access matrix model of security described in many OS texts [67, 63]. 9 More complicated sandboxes can be created using Protection Domains. These result in multiple sandboxes where the sandbox differs depending on where the code resides. Security in Java depends on correct implementation of the safety mechanisms as well as an appropriate policy. This is by no ....

Mukesh Singhal and Niranjan G. Shivaratri. Advanced Concepts in Operating Systems. McGraw Hill Inc, New York, 1994.

....characteristics of synchronous and asynchronous models are very different. There are problems that may be solved efficiently in the synchronous model when compared to the asynchronous model. Moreover, there are problems that can be solved in the synchronous model but not in the asynchronous model [56,65]. Synchronous models are easier to program and debug. However, they cannot completely exploit the inherent temporal parallelism available in applications. On the other hand, in asynchronous model of distributed computation, debugging is complicated due to total randomness in 5 the occurrence of ....

Singhal, M., and Shivaratri, N. G. Advanced Concepts in Operating Systems. McGraw-Hill, 1994.

....can be used effectively in this area as an alternative design approach that is also language independent. 3. System Components The design of operating systems has been traditionally based either on the collective kernel structure (functional decomposition) or on the objectoriented model [22]. Examples of collective kernel structures include Mach [2] and Chorus [21] Examples of object oriented operating systems include Amoeba [23, 24] and Choice [3] To our knowledge, there has so far been no proposal to address operating system design based on an aspect oriented approach. 3.1 The ....

Singhal, M., N. G. Shivaratri. Advanced Concepts in Operating Systems, McGraw-Hill, New York, 1994.

....and programming paradigms. In particular, the assignment and schedule of the processors play a key role. Both of these problems are NP complete. Many approaches are discussed in literature addressing (dynamic) run time scheduling, static) compile time scheduling and the assignment problem [16, 14, 26]. Assuming that a program has already been written (or parallelized) using parallel programming primitives of a language or development environment, our approach consists in redistributing or migrating code from one processor to another one at run time, based on dynamic, adaptive and distributed ....

M. Shinghal and N.G. Shivaratri. Advanced Concepts in Operating Systems. MacGraw Hill, 1994. 14

.... on mobile distributed computing environment include causal ordering of messages [2, 15, 22] fault tolerance [1] distributed snapshot [21] and distributed checkpointing [9, 16] In this paper, we study one fundamental problem in distributed computing, the termination detection problem [17]. A distributed system can be generally defined as a set of autonomous processes which cooperate with each other to complete a job. To coordinate computation and exchange data, processes may communicate with each other by message passing. Termination detection refers to the necessity of ....

....assembly pipeline, perform production management, check inventory, perform customer service, or together solve a problem. In addition, the terminated status of a distributed system is among the stable states (such as deadlock, token loss, etc. that are desired to know by system administration [17]. In the literature, termination detection schemes have also been used to help solving other distributed computing problems, such as deadlock detection [7] checkpointing [16] and crash recovery [20] There already exist two types of termination detection protocols which were designed for ....

[Article contains additional citation context not shown here]

M. Singhal and N. Shivaratri. Advanced Concepts in Operating Systems. McGraw-Hill, 1994.

....migration from Puma to Linux is possible. As the remaining underlying functionality is added to the system, the migration mechanisms will be extended commensurately. In particular, we will add a markerbased communication forwarding mechanism (cf. the Chandy Lamport global state recording algorithm [13]) as well as investigating methods to reduce migration time through analysis of address space characteristics (e.g. empty pages) and through on the fly compression. Acknowledgements The authors are indebted to the Puma team at the MPCRL and the University of New Mexico who worked on the Uniform ....

Mukesh Singhal, Niranjan G. Shivaratri. Advanced Concepts in Operating Systems. McGraw Hill, USA, 1994.

.... on mobile distributed computing environment include causal ordering of messages [2, 19, 26] fault tolerance [1] distributed snapshot [25] and distributed checkpointing [13, 20] In this paper, we study one fundamental problem in distributed computing, the termination detection problem [21]. A distributed system can be generally defined as a set of autonomous processes which cooperate with each other to complete a job. To coordinate computation and exchange data, processes may communicate with each other by messagepassing. Termination detection refers to the necessity of determining ....

....assembly pipeline, perform production management, check inventory, perform customer service, or together solve a problem. In addition, the terminated status of a distributed system is among the stable states (such as deadlock, token loss, etc. that are desired to know by system administration [21]. In the literature, termination detection schemes have also been used to help solving other distributed computing problems, such as deadlock detection [11] checkpointing [20] and crash recovery [24] In this paper, we start with investigating two types of termination detection protocols which ....

[Article contains additional citation context not shown here]

M. Singhal and N. Shivaratri. Advanced Concepts in Operating Systems. McGraw-Hill, 1994.

....systems. In this paper we describe how mutual exclusion can be implemented on pipelined optical buses. The problem of mutual exclusion is critical to the shared memory paradigm of parallel computation since it arises whenever concurrent access to shared resources by several sites is involved [4]. For correctness it is necessary to guarantee that the shared resource is accessed by only one site at a time (i.e. mutual exclusion) In addition to assuring mutual exclusion, techniques for mutual exclusion in distributed systems must exhibit the following characteristics: 1) Freedom from ....

....for d ij . 3.3 Performance Analysis In this section, we analyze the performance of the Timestamp algorithm of the previous section assuming the folded bus architecture. We begin by defining some performance measures for mutual exclusion algorithms. Our definitions are adapted from those in [4]. Receive Request(TIME t, Proc ID i, Resource ID r, Time t i ) begin if (there is a resource req msg at P i s receiver at time t) assume that requests originating at P i are ignored. begin RECV(message)# Processor ID j = message.getProcID( # Resource ID s = message.getResourceID( # if ....

M. Singhal, N. Shivaratri, Advanced Concepts in Operating Systems. New York: McGraw-Hill, 1994.

....consistent, global framework for application support. 1. Introduction In ten or fifteen years, we bet, this workshop will become the Workshop on Hot Topics in Run Time Application Support. As a community, we used to define an operating system as hardware abstraction and resource management [32, 34]. It is the authors opinion that there are a continuum of tools that provide application support, from compiler to thread scheduler, and that the choice of which tools to apply statically and which to apply at run time in the operating system should not be frozen by the operating system design, ....

M. Singhal and N. G. Shivaratri. Advanced concepts in operating systems. McGraw-Hill, 1994.

....software fault tolerance mechanisms are added to programs based on the generative communication paradigm. Applying layer 3 methods yields an acceptable level of fault tolerant execution for such applications. Application of 5 There are more problems with BBER. An in depth treatment is given in [MN94] 7. Combining Generative Communications and Software Fault Tolerance 11 layer 3 strategies to these distributed programs is then examined in greater detail. As all inter process communication is done via the tuple space, there is already a system entity in place to keep the message log for each ....

Mukesh Singhal and Niranjan Shivaratri. Advanced Concepts in Operating Systems. Series in Computer Science. Mc Graw Hill, 1994.

....mechanisms are added to programs based on the generative communication paradigm. Applying layer 3 methods yields an acceptable level of fault tolerant execution for such applications. Application of layer 3 6 There are more problems with BBER. An in depth treatment is given in [MN94] 12 in( logic , I , a) out( logic , I ,b) A B T u p l e s p a c e Figure 4: IPC using the tuplespace strategies to these distributed programs is then examined in greater detail. As all interprocess communication is done via the tuplespace, there is already a system entity in place to keep the ....

Mukesh Singhal and Niranjan Shivaratri. Advanced Concepts in Operating Systems. Series in Computer Science. Mc Graw Hill, 1994.

....section, the node must release the lock. The algorithm from which this work is based assumes that none of the nodes fail. 2. 2 Requirements of Distributed Mutual Exclusion Algorithms Before describing this particular algorithm, it is helpful to remember that mutual exclusion algorithms should be[10]: ffl Free from deadlocks. Nodes should not wait for messages that will never arrive. ffl Free from starvation. A node should acquire the lock in a finite time. ffl Fair. Requests for the lock should be granted in the order they arrive in the system. ffl Fault tolerant. The algorithm should be ....

....that allocates the lock is the holder; the holder variables of all other nodes are set to this value. We define Message complexity as the number of messages sent to satisfy a request for a lock. The synchronization delay is the time between a request for the lock and entry into the critical section[10]. The inaccuracy of a node s holder variable increases the synchronization delay experienced by a requests initiated from that node. It also increases the number of messages required to satisfy a request from that node. The upper bound for the number of messages sent to request the lock is N , ....

Singhal, M. and N. Shivaratri, "Advanced Concepts in Operating Systems," McGraw-Hill, New York, 1994, p. 123.

....[LFS93, PL96, KRS96] to the fault tolerance problem on workstation clusters are based on total or partial recovery from the previous checkpointed states of applications. A simple yet efficient fault tolerant model that applies the idea of replicated execution is presented. Replicated execution [SS94b] provides fault tolerance by having a number of processes running the same program concurrently. The concept of using replicated execution for fault tolerance have been applied to distributed operating systems [Ng90] and distributed systems [OL88, MPS89] In this section, replicated execution is ....

M. Singhal and N.G. Shivaratri. Advanced Concepts in Operating Systems. New York: McGraw-Hill, 1994. (Cited on page 100).

....method nets can be obtained by means of a static analysis. Both activities produce a socalled object resource net which is a net according to the OCoN formalism but resembles the resource allocation graphs used in algorithms for classical as well as distributed deadlock detection methods (cf. [15]) to some extent. If the potential for deadlock is found but cannot proved fully, the remaining potentially insecure methods can be put into the focus of a permanent or on demand deadlock detection at runtime. Moreover, the restrictions keep the developer from introducing unwanted dependencies by ....

M. Shinghal and N.G. Shivaratri. Advanced Concepts in Operating System. Mc Graw-Hill, Inc., 1995.

....can be ignored for further processing and physically discarded from both own history and out histories. The problem of choosing the time moment for the clipping to be performed is a well known problem is distributed systems that of establishing a consistent cut within the system s execution[16]. The algorithm we adopt is the gossip message technique [17] which offers an asynchronous broadcast facility with a causal order of delivery, allowing the information to disseminate within the system at arbitrary times. In the gossip technique, beside functional messages, the processes also ....

....then processing may continue. The criteria for deciding if a security error is recoverable 1 or not are strictly dependent on the application under consideration and its security specifications. Also depending on the application and particular error detected, either forward or backward recovery [16] may be used for treating a recoverable error. As backward recovery is a general mechanism, it does not bring any particular problem to run time security evaluation. The point of interest is forward recovery, when specific actions must be taken for each type of error. As errors are detected as ....

M. Singhal and N. Shiravaratri. Advanced Concepts in Operating Systems. McGraw-Hill, Inc., 1994.

....toolkit for implementing scheduling algorithms. This will include facilities for obtaining state information for the underlying system and the task suite to be scheduled, as well as expressing various portions of a scheduling policy (the transfer, location, selection, and information policies [16]) This will be flexible enough to implement all classes of scheduling algorithms from the existing literature, and will be extensible so as to accommodate future algorithms. The toolkit will also facilitate implementation of local scheduling policies. a representation of the underlying ....

M. Singhal and N. G. Shivaratri, Advanced Concepts in Operating Systems, McGraw-- Hill, 1994, ISBN 0-07-13668-1.

....must be managed to avoid overtaking storage memory resources. A control algorithm must be in place to handle the possibly constant flow of metrics that are collected and applied to the interface learning nodes. This problem is not unlike the process control problem operating systems designers face [11]. The control algorithm must ensure the interface information nodes are accessed and updated in a deterministic fashion to ensure accuracy of the interface s reasoning. 8 Results and Conclusions Constructing a reasoning architecture into a user interface in order to facilitate intelligent ....

Mukesh Singhal and Niranjan G. Shivaratri. Advanced Concepts in Operating Systems. McGraw-Hill, Inc, New York, NY, 1994.

No context found.

M. Singhal, N Shivara t ri, Advanced Concepts in Operating Systems, Mc Graw Hill 1994. p.98. 

No context found.

M. Singhal and N. G. Shivaratri. Advanced Concepts in Operating Systems. McGraw--Hill, 1994. ISBN 0-07-13668-1.

No context found.

M. Singhal and N. G. Shivaratri. Advanced Concepts in Operating Systems. McGraw--Hill, 1994. ISBN 0-07-13668-1.

No context found.

Mukesh Singhal and Niranjan Shivaratri, Advanced Concepts in Operating Systems, McGraw Hill, 1994

No context found.

M. SINGHAL AND N. G. SHIVARATRI. Advanced Concepts in Operating Systems. McGraw-Hill, 1994.

No context found.

M. Singhal and N. Shivaratri. Advanced Concepts in Operating Systems. McGraw-Hill, 1994. ISBN 0-07-057572-X.

No context found.

Singhal, M., Shivaratri, N.G., Advanced Concepts in Operating Systems, McGraw Hill, Inc., New York, 1994.

First 50 documents

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