A heterogeneous distributeddatabase system (HDDS) integrates existing databasesystems to support global transactions which access multiple databases, where local databasesystems may use different concurrency contro...
详细信息
Replication is a key factor for improving the availability of data in a distributed system. A major restriction in using replication is that replicated copies must behave like a single copy;i.e. mutual and intemal con...
详细信息
An atomic commit protocol can ensure that all participants in a distributed transaction reach consistent states, whether or not system or network failures occur. One widely used protocol is the two-phase commit (2PC) ...
详细信息
ISBN:
(纸本)0818635703
An atomic commit protocol can ensure that all participants in a distributed transaction reach consistent states, whether or not system or network failures occur. One widely used protocol is the two-phase commit (2PC) protocol, which has long appeared in the literature. Much of the literature focuses on improving performance in failure cases by providing a non blocking 2PC that streamlines recovery processing at the expense of extra processing in the normal case. We focus on improving performance in the normal case based on two assumptions: first that networks and systems are becoming increasingly reliable, and second that the need to support high-volume transactions requires a streamlined protocol for the normal case. In this paper, various optimizations are analyzed in terms of reliability, savings in log writes and network traffic, and reduction in resource lock time. Its unique contributions include the description of some optimizations not described elsewhere and a systematic comparison of the optimizations and the environments where they cause the most benefit.
The paper develops a methodology for implementing semantic synchronization specifications with ordinary Read (Share) and Write (Exclusive) locks, and it describes a software system SLEVE that implements the methodolog...
详细信息
ISBN:
(纸本)0818635703
The paper develops a methodology for implementing semantic synchronization specifications with ordinary Read (Share) and Write (Exclusive) locks, and it describes a software system SLEVE that implements the methodology. An application developer who defines a conflict relation over an arbitrary set of synchronization primitives, such as abstract type operations, events, or semantic lock types, can use SLEVE to implement the specification on an underlying system that supports only Read/Write locking. The methodology can be used to synchronize either transactions or processes, and it can be used in multiversioning schemes to control version proliferation.
In this paper, a number of mechanisms are described that can be used in the design of a protocol converter for authentication and key distribution protocols. First, the scope of the mechanisms is defined. we mark out ...
详细信息
ISBN:
(纸本)0818633700
In this paper, a number of mechanisms are described that can be used in the design of a protocol converter for authentication and key distribution protocols. First, the scope of the mechanisms is defined. we mark out the class of authentication systems that were considered during the design of the mechanisms. A first mechanism, based on proxies and a synchronization protocol, allows for a transparent protocol conversion. It is generic, and can be tailored to different specific situations. The second mechanism addresses the problem of the statefulness of the protocol convertor. Both mechanisms can be used separately or in combination. When properly combined, they provide for a robust, transparent, and safe protocol convertor for authentication and key distribution protocols.
The OSI directory system is a distributed directory information database which incorporates distributed algorithms for name resolution and search to support applications such as electronic mail, White Pages, authentic...
详细信息
ISBN:
(纸本)0818637706
The OSI directory system is a distributed directory information database which incorporates distributed algorithms for name resolution and search to support applications such as electronic mail, White Pages, authentication services, etc. To improve the performance of requests for name resolution and search, it is necessary to lower the cost of name lookups for nonlocal names. The OSI directory system and other distributed naming systems use name caching to minimize such costs. In this paper, we evaluate the impact of name caching on the performance of the OSI directory system. We consider the issues of cache sizing and cache replacement policies. We find that a locality of reference property holds in name resolution requests, and hence name caching does increase performance significantly. Using trace-driven simulation, we show that small caches (smaller than 30 items) yield hit ratios up to 60% and decrease the average name resolution time by 60%. For small caches, the LRU (Least Recently Used) replacement policy is better than other implementable policies. Large caches yield predictably larger hit ratios. For large caches, however, the LRU policy is not better than a random replacement policy. We also find that partitioning the cache buffer into a small number of independent caches, each one associated with a different kind of name requests, further decreases the average name resolution time.
Query optimization in heterogeneous databasesystems is not always possible since the component DBMS may not have the ability to transmit necessary information. However, these systems need query optimization because t...
详细信息
ISBN:
(纸本)0897915585
Query optimization in heterogeneous databasesystems is not always possible since the component DBMS may not have the ability to transmit necessary information. However, these systems need query optimization because the cost of transmitting large quantities of data across diverse databases is very high. We propose a query strategy which uses hypothesis testing to determine which of two sets of data are larger. Our experiments show that this strategy is very likely to select the smaller set when the sampling results fall outside a region of uncertainty we call the `grey zone.' This provides query optimization without transmission of database statistics.
In this paper we consider the problem of update propagation in a multidatabase environment. Initially, the update propagation strategy is discussed in terms of minimization of the total communication cost of such oper...
详细信息
ISBN:
(纸本)0780312333
In this paper we consider the problem of update propagation in a multidatabase environment. Initially, the update propagation strategy is discussed in terms of minimization of the total communication cost of such operation. We indicate that the Steiner Tree is a more suitable model of update propagation than a minimum spanning tree approach. Since the Steiner Tree problem is NP-hard, the update propagation require an application of heuristic method. Subsequently, further optimization in respect of system's response time is considered and an algorithm for computation of a propagation strategy optimal with respect of this criterion is presented and analyzed. The update propagation response time can be improved by selecting a propagation center in the network;an appropriate algorithm is presented.
Current techniques for query scheduling in a parallel database system schedule a single query at a time. This paper investigates scheduling of queries for parallel databasesystems by dividing the workload into batche...
详细信息
ISBN:
(纸本)0818635703
Current techniques for query scheduling in a parallel database system schedule a single query at a time. This paper investigates scheduling of queries for parallel databasesystems by dividing the workload into batches. We propose scheduling algorithms which exploit the common operations within the queries in a batch. The performance of the proposed algorithms is studied using a simple analytical model and a detailed simulation model. We show that batch scheduling can provide significant savings compared to single query scheduling for a variety of system and workload parameters.
Our project is studying the process by which groups of individuals work together to build large, complex structures of ideas and developing a distributed hypermedia collaboration environment (called ABC) to support th...
详细信息
ISBN:
(纸本)0897916247
Our project is studying the process by which groups of individuals work together to build large, complex structures of ideas and developing a distributed hypermedia collaboration environment (called ABC) to support that process. This paper focuses on the architecture and implementation of the distributed Graph Storage (DGS) component of ABC. The DGS supports a graph-based model, conservatively extended to meet hypermedia requirements. Some important issues addressed in the system include scale, performance, concurrency semantics, access protection, location independence, and replication (for fault tolerance).
暂无评论