臺灣博碩士論文加值系統

近年來，先決資料庫系統已經證實為在絕大多數的資料可以放進記憶體的情況下，可於一般商用電腦組成之叢集系統中達到高吞吐量的效能，並同時保證多個資料庫副本間能夠達到高一致性。然而，一般的線上交易處理工作的資料量通常不容易全部放進一般商用電腦的記憶體中。緩衝區置換所造成的長時間執行交易會對先決資料庫系統的吞吐量有關鍵性的影響。在這篇論文中，我們提出名為IDA的架構。IDA重新設計了交易執行的架構並減少長時間執行交易的出現量與衝擊性。在IDA中，緩衝區層使用了附帶版本號並以紀錄為基礎的緩衝區池，並消除了並行控制的必要性。IDA的儲存層被重新設計以充分利用連續順序寫入的好處並簡化復原層。我們在資料庫系統上實做了IDA，並以實驗證實IDA的效果。

The deterministic database systems have recently been shown to yield high throughput on a cluster of commodity machines while ensuring the strong consistency between replicas, provided that most of the data can be fit in the memory on these machines. However, standard OLTP data is unlikely to fit in the memory of commodity machines. The long transactions induced by the buffer swapping have a critical impact to the throughput of deterministic database systems. In this paper, we present IDA, a redesigned transaction execution architecture which reduces the occurrence and impact of long transactions. IDA redesigns the buffer layer with versioned record-based buffer pool and eliminates the need of concurrency control layer. The storage layer is redesigned to take advantage of sequential writes and simplify the recovery layer. We implement IDA on a real database system. Experiments are conducted and the results demonstrate the effectiveness of IDA.

先決式資料庫的逆向存取方法 ii
Inversion of Data Access for Deterministic Databases iii
Acknowledgments iv
Abstract v
摘要 vi
Table of Contents vii
List of Figures x
List of Tables xi
Chapter 1 Introduction 1
Chapter 2 Background and Challenges 5
2.1. Deterministic Database Systems 5
2.2. The Impact of Long Transactions 6
Chapter 3 Inversion of Data Access 7
3.1. Scheduling Layer 7
3.2. Buffer Layer 7
3.2.1. Concurrency Control 8
3.3. Storage Layer 9
3.3.1. Recovery 10
Chapter 4 Experiments 11
4.1. Microbenchmark 11
4.1.1. Partitioning T-Graph in Real-Time 11
4.1.2. The Impact of Long Read Transactions 12
4.1.3. The Effect of Read-Write Transaction Ratio 12
Chapter 5 Related Work 14
5.1. Concurrency Control in Deterministic Database Systems 14
5.2. Optimistic Concurrency Control 14
Chapter 6 Conclusion 16
Bibliography 17

[1] Alexander Thomson and Daniel J Abadi. The case for determinism in database systems. Proc. of the VLDB Endowment, 3(1-2):70-80, 2010.
[2] Alexander Thomson, Thaddeus Diamond, Shu-Chun Weng, Kun Ren, Philip Shao, and Daniel J Abadi. Calvin: Fast distributed transactions for partitioned database systems. In Proc. of SIGMOD, pages 1-12. ACM, 2012.
[3] C. Mohan, B. G. Lindsay, and R. Obermarck. Transaction management in the r* distributed database management system. ACM Trans. Database Syst.,1986.
[4] Kun Ren, Alexander Thomson, and Daniel J. Abadi. Lightweight Locking for Main Memory Database Systems. In PVLDB 6(2): 145-156, 2012.
[5] H.T. Kung, John T. Robinson. On optimistic methods for concurrency control. ACM Transactions on Database Systems (TODS), TODS Volume 6 Issue 2, June 1981 Pages 213-226