Seen hierarchically, an information network can have 3 different functions associated with it: 1. information processing, 2. network processing, and 3. database function. A distributedcomputing system is consider...
详细信息
Seen hierarchically, an information network can have 3 different functions associated with it: 1. information processing, 2. network processing, and 3. database function. A distributedcomputing system is considered mainly at the network processing level in the context of load-leveling, a process by which a given workload is spread evenly over the processors in a distributed system. Assuming a horizontally distributed system, formulations of the edge-failure and node-failure recovery problems from the standpoint of load-leveling are developed. The conditions for the existence of solutions to these problems are analyzed, and simple algorithms are presented for these problems. In connection with the node-failure recovery problem, the concept of a node-failure metric to characterize different possible solutions is introduced, exploiting the idea of the strength of processors. A possible application of the recovery methods in the context of reconfiguration of distributed database systems is examined.
A graph matching approach is proposed in this paper for solving the task assignment problem encountered in distributed computing systems. A cost function defined in terms of a single unit, time, is proposed for evalua...
详细信息
A graph matching approach is proposed in this paper for solving the task assignment problem encountered in distributed computing systems. A cost function defined in terms of a single unit, time, is proposed for evaluating the effectiveness of task assignment. This cost function represents the maximum time for a task to complete module execution and communication in all the processors. A new optimization criterion, called the minimax criterion, is also proposed, based on which both minimization of interprocessor communication and balance of processor loading can be achieved. The proposed approach allows various system constraints to be included for consideration. With the proposed cost function and the minimax criterion, optimal task assignment is defined. Graphs are then used to represent the module relationship of a given task and the processor structure of a distributedcomputing system. Module assignment to system processors is transformed into a type of graph matching, called weak homomorphism. The search of optimal weak homomorphism corresponding to optimal task assignment is next formulated as a state-space search problem. It is then solved by the well-known A* algorithm in artificial intelligence after proper heuristic information for speeding up the search is suggested. An illustrative example and some experimental results are also included to show the effectiveness of the heuristic search.
Onboard spacecraft computing system is a case of a functionally distributed system that requires continuous interaction among the nodes to control the operations at different nodes. A simple and reliable protocol is d...
详细信息
Onboard spacecraft computing system is a case of a functionally distributed system that requires continuous interaction among the nodes to control the operations at different nodes. A simple and reliable protocol is desired for such an application. This paper discusses a formal approach to specify the computing system with respect to some important issues encountered in the design and development of a protocol for the onboard distributed system. The issues considered in this paper are concurrency, exclusiveness and sequencing relationships among the various processes at different nodes. A 6-tuple model is developed for the precise specification of the system. The model also enables us to check the consistency of specification and deadlock caused due to improper specification. An example is given to illustrate the use of the proposed methodology for a typical spacecraft configuration. Although the theory is motivated by a specific application the same may be applied to other distributedcomputing system such as those encountered in process control industries, power plant control and other similar environments.
With the advent of distributed computing systems, the problem of deadlock, which has been essentially solved for centralized computingsystems, has reappeared. Existing centralized deadlock detection techniques are ei...
详细信息
With the advent of distributed computing systems, the problem of deadlock, which has been essentially solved for centralized computingsystems, has reappeared. Existing centralized deadlock detection techniques are either too expensive or they do not work correctly in distributed computing systems. Although several algorithms have been developed specifically for distributedsystems, the majority of them have also been shown to be inefficient or incorrect. Additionally, although fault-tolerance is usually listed as an advantage of distributed computing systems, little has been done to analyze the fault tolerance of these algorithms. This thesis analyzes four published deadlock detection algorithms for distributed computing systems with respect to their performance in the presence of certain faults. A new deadlock detection algorithm is then proposed whose efficiency and fault tolerance are adjustable.
A distributed system - a special computer network with a high degree of cohesiveness, transparency, and autonomy - has such potential benefits as: 1. high availability, 2. greater reliability, 3. improved throughput a...
详细信息
A distributed system - a special computer network with a high degree of cohesiveness, transparency, and autonomy - has such potential benefits as: 1. high availability, 2. greater reliability, 3. improved throughput and response time, 4. resource sharing and load leveling, and 5. modular and incremental growing. This research presents the framework and model of a network operating system (NOS) called Multicomputer Integrator Kernel (MIKE), for use in distributedsystems, particularly in the distributed Double-Loop Computer Network (DDLCN). MIKE provides system-transparent operation for users and maintains cooperative autonomy among local hosts. MIKE is based on the object model and a unique ''task'' concept and uses message passing as an underlying semantic structure. The distributed system kernel has a layered protocol to support NOS, providing a flexible organization in which system-transparent resource sharing and distributedcomputing can evolve in a modular fashion. The NOS model and the notion of ''task'' are presented, and the system naming convention is examined. A 2-level process interaction model is also described. The protection mechanism is discussed, emphasizing maximal error confinement. A scenario for system-transparent resource sharing utilizing these concepts is given. A multilayer, multidestination protocol structure is outlined.
After a failure occurs in a distributedcomputing system, it is often necessary to reorganize the active nodes so that they can continue to perform a useful task. The first step in such a reorganization or reconfigura...
详细信息
After a failure occurs in a distributedcomputing system, it is often necessary to reorganize the active nodes so that they can continue to perform a useful task. The first step in such a reorganization or reconfiguration is to elect a coordinator node to manage the operation. This paper discusses such elections and reorganizations. Two types of reasonable failure environments are studied. For each environment assertions which define the meaning of an election are presented. An election algorithm which satisfies the assertions is presented for each environment.
A distributedcomputing system is characterized as follows: 1. It has a number of hardware processors with each processor having a certain computing capability. 2. Communications links exist between or among the proce...
详细信息
A distributedcomputing system is characterized as follows: 1. It has a number of hardware processors with each processor having a certain computing capability. 2. Communications links exist between or among the processors. 3. A number of functional components reside on each processor, and the software components of different processors may be synchronous or asynchronous. 4. The functional components interact for the purpose of performing some system-wide functions. To deal with a distributedcomputing system as defined, the framework of an approach is presented for developing the system's design specifications. In this approach, data components and functional components are handled separately. A notation for the design is supplied, and performance estimates are given.
To answer the question contained in the above title we start with the examination of properties of both the “classical” computer systems, the batch computer system and the process control computer system. Distribute...
详细信息
To answer the question contained in the above title we start with the examination of properties of both the “classical” computer systems, the batch computer system and the process control computer system. distributed computing systems show new properties, introduced by the computing system communication network as well as by novel man-system communication means and methods. These new properties influence both types of distributed computing systems in such a way that their properties become similar and differences degrade to parameter variations.
暂无评论