Amortized efficiency of list update and paging rules
Communications of the ACM
An adaptive data replication algorithm
ACM Transactions on Database Systems (TODS)
Competitive Analysis of Caching in Distributed Databases
IEEE Transactions on Parallel and Distributed Systems
Online computation and competitive analysis
Online computation and competitive analysis
Principles of distributed database systems (2nd ed.)
Principles of distributed database systems (2nd ed.)
Stable and fault-tolerant object allocation
Proceedings of the nineteenth annual ACM symposium on Principles of distributed computing
Optimal Placement of Replicas in Trees with Read, Write, and Storage Costs
IEEE Transactions on Parallel and Distributed Systems
Distributed Systems for System Architects
Distributed Systems for System Architects
Distributed Systems: Principles and Paradigms
Distributed Systems: Principles and Paradigms
Popularity-Aware Greedy Dual-Size Web Proxy Caching Algorithms
ICDCS '00 Proceedings of the The 20th International Conference on Distributed Computing Systems ( ICDCS 2000)
Code Cache Management Schemes for Dynamic Optimizers
INTERACT '02 Proceedings of the Sixth Annual Workshop on Interaction between Compilers and Computer Architectures
Buffer Cache Management: Predicting the Future from the Past
ISPAN '02 Proceedings of the 2002 International Symposium on Parallel Architectures, Algorithms and Networks
| Hi-index | 0.01 |
In this paper, we address the performance of distributed database systems with buffer constraints. Specifically, our objective is to design and analyze efficient data allocation and replication strategies to minimize the total servicing cost for an arbitrary read/write request sequence, under finite buffer constraints of the nodes in the system. When the available buffer space in a node is not enough to store a copy of an object, the decision has to be made on whether or not we should evict one or more objects in use to give room for the new object copy. In this paper, we design and analyze the data replication strategies with the model of Dynamic Window Mechanism (DWM) algorithm jointly implemented with different types of object replacement strategies (No Replacement, LRU, and LFU) commonly found in practice. We consider situations wherein the object sizes are identical as well as heterogeneous. We will show the impact on the performance of the allocation and replication strategies due to the limited local database buffer capacities. We analyze and quantify theoretically (using competitive analysis) the performances of all the proposed algorithms. Further, we perform rigorous simulation experiments to validate the findings with respect to several influencing parameters. Several useful conclusions are drawn based on the experimental results and we highlight the usefulness of the algorithms under different situations.