DESIGN AND ANALYSIS OF OBJECT ALLOCATION AND REPLICATION ALGORITHMS IN DISTRIBUTED DATABASES FOR STATIONARY AND MOBILE COMPUTING SYSTEMS LIN WUJUAN (B.Eng., Xi’an Jiaotong University, PRC ) A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ELECTRICAL & COMPUTER ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2004 To Parents and Wife. i Acknowledgements Firstly, I am greatly indebted to my mentor, Assistant Professor Bharadwaj Veeravalli, for all the supports, valuable suggestions and insightful comments that made this work possible. It was a pleasant and challenging time working with him for the past three years, while he incessantly and persuasively imparted me a lot on doing research. I benefited much from his valuable critiques and rigorous research attitude. Secondly, I would like to take this opportunity to express my deepest appreciation to my wife, Hu Xiaohong, for her selfless love, endless patience, understanding, and encouragement provided throughout the long duration of my research work. Words alone cannot convey my gratefulness to my beloved parents, brother, and sisters for their continuous encouragement and supports throughout my life. Without them, I could not come so far in my long study life. My heartfelt thanks to the National University of Singapore (NUS) for granting me research scholarship and the Open Source Software Laboratory (OSSL) for providing me all the facilities. Special thanks to all my friends in OSSL for creating a conducive and joyful studying and working ambience, making my study and life in NUS fruitful and enjoyable. Finally, I would like to pass my gratitude to all those who have directly or indirectly helped me during the course of my research with their ideas, inputs or moral support. ii Contents List of Figures List of Tables Summary Introduction v vii ix 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Issues to Be Studied and Main Contributions . . . . . . . . . . . . . . . . . . . 1.3 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 System Modeling 11 2.1 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.2 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Object Management in Stationary Computing Environments 16 iii 3.1 3.2 3.3 3.4 Preliminaries and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . 17 3.1.1 SA Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.1.2 DA Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 DWM Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.2.1 Cost Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.2.2 Window Mechanism of DWM Algorithm . . . . . . . . . . . . . . . . . 27 3.2.3 Servicing of Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.4 Competitive Analysis of DWM Algorithm . . . . . . . . . . . . . . . . 34 ADRW Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.3.1 Cost Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.3.2 Distributed Request Window Mechanism . . . . . . . . . . . . . . . . . 49 3.3.3 Competitive Analysis of ADRW Algorithm . . . . . . . . . . . . . . . . 54 3.3.4 Failure and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Object Management in Mobile Computing Environments 4.1 63 DWM Algorithm in MCEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.1.1 Cost Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.1.2 Servicing of Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.1.3 Competitive Analysis of DWM Algorithm . . . . . . . . . . . . . . . . 66 iv 4.2 4.3 4.4 RDDWM Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.2.1 Cost Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4.2.2 Window Mechanism of RDDWM Algorithm . . . . . . . . . . . . . . . 77 4.2.3 Servicing of Request Sub-sequences . . . . . . . . . . . . . . . . . . . . 80 4.2.4 Competitive Analysis of RDDWM Algorithm . . . . . . . . . . . . . . 81 4.2.5 Simulation Results and Discussions . . . . . . . . . . . . . . . . . . . . 83 ADRW Algorithm in a MCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 4.3.1 Cost Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 4.3.2 Distributed Request Window Mechanism . . . . . . . . . . . . . . . . . 87 4.3.3 Competitive Analysis of ADRW Algorithm . . . . . . . . . . . . . . . . 88 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Experiments with ADRW Algorithm 93 5.1 Experimental System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 5.2 Experimental Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . 95 5.3 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Conclusions and Future Work 107 Bibliography 112 Author’s Publications 121 v List of Figures 2.1 An illustration of the system model of a DDBS . . . . . . . . . . . . . . . . . 11 3.1 Illustration of the concurrent control mechanism . . . . . . . . . . . . . . . . . 25 3.2 Illustration of phase partition in DWM algorithm – Heuristic . . . . . . . . . 29 3.3 Example of the working policy of the window mechanism in DWM algorithm . 29 3.4 Illustration of two extreme cases in DWM algorithm . . . . . . . . . . . . . . . 31 3.5 Competitive ratio comparison of DWM, DA, and SA algorithm in the SCE . . 45 3.6 Illustration of the TEN policy in server pj for a non-data-processor pi . . . . . 52 3.7 Illustration of the TEX policy in a data-processor pi . . . . . . . . . . . . . . . 54 3.8 Illustration of Phase Partition technique – Heuristic . . . . . . . . . . . . . . 55 4.1 Performance comparison of RDDWM under short deadline periods and sufficiently long deadline periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 5.1 84 Performance comparison of RDDWM under random deadline periods (between [1,10] time units) and sufficiently long deadline periods . . . . . . . . . . . . . 85 Logical network topology of the experimental system . . . . . . . . . . . . . . 94 vi 5.2 Cost performance of ADRW, SA, and DA algorithm when the request window size k = 10 in ADRW algorithm and each node has the same probability of read/write request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 98 Cost performance of ADRW, SA, and DA algorithm when the request window size k = 10 in ADRW algorithm and each node has different probability of read/write request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5.4 Cost performance of ADRW, SA, and DA algorithm when the request window size k = 30 in ADRW algorithm and each node has the same probability of read/write request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 5.5 Cost performance of ADRW, SA, and DA algorithm when the request window size k = 50 in ADRW algorithm and each node has the same probability of read/write request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 5.6 Number of request window transferring in ADRW algorithm when each node has the same probability of read/write request and k=10, 30, and 50 . . . . . . 104 5.7 Average cost for servicing a request when each node has the same probability of read/write request and the request window size k=10, 30, and 50 in ADRW algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 vii List of Tables 2.1 Glossary of Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.1 The adjustment of Ao when DA algorithm services σo . . . . . . . . . . . . . . 23 3.2 Window mechanism of DWM algorithm . . . . . . . . . . . . . . . . . . . . . . 28 3.3 Test-and-Enter (TEN) policy . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.4 Test-and-Exit (TEX) policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.1 Window mechanism of RDDWM algorithm . . . . . . . . . . . . . . . . . . . . 78 4.2 Competitive ratios of SA, DA, DWM, RDDWM, and ADRW algorithm in both the SCE and the MCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.1 Hardware configurations of the experimental system . . . . . . . . . . . . . . . 94 5.2 Mean request arriving interval at each node . . . . . . . . . . . . . . . . . . . 96 5.3 Results of the experiments when the request window size k = 10 in ADRW algorithm and each node has the same probability of read/write request . . . . 5.4 97 Probability of read request at each node . . . . . . . . . . . . . . . . . . . . . 100 5.5 Results of the experiments when the request window size k = 30 in ADRW algorithm and each node has the same probability of read/write request . . . . 102 5.6 Results of the experiments when the request window size k = 50 in ADRW algorithm and each node has the same probability of read/write request . . . . 102 107 Chapter Conclusions and Future Work In this thesis, we have addressed two important issues, object allocation and object replication, for the object management process (OMP) in DDBSs. Our emphasis was focused on design and analysis of strategies to minimize the associated (monetary) cost of servicing requests that arrive at a DDBS. As mentioned in Chapter 1, we considered two distinct application domains that are in place - Stationary Computing Environment (SCE) domain and Mobile Computing Environment (MCE) domain. In Chapter 3, for the application domains of SCE, we first presented mathematical cost models that consider all the costs (I/O cost, control-message/data-message transferring cost) associated with the servicing of a request, and then proposed two dynamic on-line object allocation and replication algorithms, DWM algorithm (for centralized control DDBSs) and ADRW algorithm (for decentralized control DDBSs). Based on different conditions, these two dynamic algorithms were carefully designed to adjust the object allocation schemes so as to minimize the total servicing cost of current request sequence (DWM) or future expected request sequence (ADRW). Using competitive analysis, we rigorously evaluated the performance of these two algorithms. Chapter Conclusions and Future Work 108 Further, we extended our design and analysis of the above two algorithms to the application domains of MCE in Chapter 4. For the DWM algorithm, we first modified its cost model proposed in SCEs to suit the conditions of MCEs, and carried out similar competitive analysis as that in SCEs. In addition, we modified the DWM algorithm to the RDDWM (Real-time Decentralized Dynamic Window Mechanism) algorithm to handle the real-time requests in a RTDDBS. Competitive analysis was carried out to quantify the performance of RDDWM algorithm under two different extreme conditions, i.e., when the deadline periods of all the requests are sufficiently long and when the deadline periods of all the requests are very short. We also conducted a simulation study to capture the performance of RDDWM algorithm with respect to different deadline periods, different number of requests, and different read/write request probability. Similarly, for the ADRW algorithm, we discussed on how this algorithm can be adopted in a MCE and also evaluated its competitive property under two different conditions. From the analysis results, we realized that algorithms designed for a SCE can be tuned to suit the conditions of a MCE, and the competitive feature of an on-line algorithm may vary when the system environments change. Furthermore, we carried out experiments to study the performance of ADRW algorithm under several influencing conditions in a SCE. We conducted detailed performance analysis by comparing the ADRW algorithm with SA (Static Allocation) and DA (Dynamic Allocation) algorithm [73]. The experimental results, coupled with the competitive analysis results presented in Chapters and 4, clearly demonstrated the influences of several relevant parameters on the system performance, and gave more insights on the design of object allocation and replication algorithms for DDBSs. Noteworthily, in order to satisfy the t-availability constraint of an object o in the system, there is a server set S(o) (|S(o)| = t) in each of the algorithms presented in this thesis. Though obtaining an optimal server set S(o) for an object o in a general network may be a Chapter Conclusions and Future Work 109 NP-complete problem [16, 37, 78], system designers are expected to choose those processors, which have larger processing capacities and/or higher reliability in the system, as the servers of an object. Indeed, this is an implementation level issue to be considered, as done in our experiments. Although the major focus in this thesis was on designing object allocation and replication algorithms in DDBSs, the concepts and issues seem to be applicable to a wide variety of application domains. For example, the DWM algorithm can be easily presented in a video-on-reservation (VOR) system [12, 68, 79], where the video warehouses (VWHs) can be considered as the server set of objects (objects in this case are multimedia documents) and the Personalized Service Agents (PSAs) [35] form a centralized control unit (CCU) that collects requests for a certain time period and then serve these requests to deliver a near-optimal performance. As a second example, the ADRW algorithm can be adopted to solve Web Caching problems on WWW, where each web proxy [9, 44, 71] represents a node in the network and a web server can be considered as an object (web-page) server. The web server can then decide whether or not a web-page should be replicated in a web proxy based on the TEN policy of ADRW algorithm. Similarly, a web proxy can also decide whether or not a web-page should be discarded based on the TEX policy of ADRW algorithm. We shall now present several issues that are beyond the scope of this thesis and can serve as our future work. First of all, when the size of requested objects in a DDBS is large, one of the unavoidable issues to be addressed is the storage space availability in the network [2, 6, 33, 67]. Applications, such as VOD or VOR systems, demand a very large storage infrastructure, especially when handling movie-on-demand requests. The size of an object in this case may be rather large (100s of Mega-Bytes or several Giga-bytes magnitude), and hence the limited storage space availability becomes a bottleneck for superior system performance. For instance, a multimedia Chapter Conclusions and Future Work 110 object stored in a node may be evicted to reserve sufficient space for another object that will be replicated in the node. In such a scenario, efficient memory replacement algorithms such as LRU (Least Recently Used), LFU (Least Frequently Used), FIFO (First-In First-Out) and LIFO (Last-In First-Out), will play a key role [28, 34, 42, 43, 48]. Therefore, when the memory capacity is taken into account, it becomes more meaningful and challenging for us to integrate efficient memory replacement algorithms into the object allocation and replication algorithms. Further, in this thesis, we have exclusively considered an object as an indivisible entity such as a multimedia document, a web-page, or a small text file, as per the underlying application domain. A complete object model Common Object Request Broker Architecture (CORBA) is built in [50], including object types, parameters, interfaces, attributes, etc. Under the CORBA standards, an object is defined as “an identifiable, encapsulated entity that provides one or more services that can be requested by a client”. This means that an entity cannot be partitioned and supplied to the requested site. However, in case when an object can be partitioned into several fragments [63, 81], a single request may be serviced by several processors concurrently, each responding for one or more fragments of the object. In such a scenario, load balancing is an additional issue to be considered [24, 25, 26, 63]. Further, the study of the case when objects are divisible illuminates us to consider a DDBS where each request arriving can demand several objects. The contents of this thesis will serve as a basis for designing competitive on-line algorithms to handle such requests. Besides the object allocation/replication issues, we have introduced another important issue in the OMP, i.e., object migration (refer to Chapter 1), which also receives an immense research interest in recent years. As far as object migration is concerned, there are several interesting open problems such as designing object migration algorithms with optimal competitive ratio for general networks or special topology networks [1, 5, 11, 17, 40, 70]. We are interested in Chapter Conclusions and Future Work 111 designing competitive object migration algorithms based on the cost models presented in this thesis. As mentioned earlier, using an on-line object management algorithm, an object in the system may be dynamically replicated in several new locations (nodes) or migrated from one location to another, as per the arriving request patterns, to improve certain system performance. Practically, the object allocation scheme (refer to Chapter 3) serves as a directory that indicates the locations of an object. It may be noted that the processes such as object replication and object migration are essentially modifying this directory. Thus, we are confronted with a problem on deciding whether this directory be made centralized, distributed, partitioned, or replicated in the system. In addition, we must also address on how the concurrency control can be taken care to maintain the consistency when the directory is updated. As a result, from the viewpoint of OMP, this object directory can itself be considered as an object in the application system. Designing efficient directory management mechanisms therefore presents a potential research area for our future work. Lastly, in this thesis, our objective is to minimize the total servicing cost of the arriving requests, and use competitive analysis to evaluate the performance of the proposed algorithms. There are several other performance metrics can be studied when designing object management algorithms in a DDBS, such as minimizing the mean response time of transactions, minimizing the complexity of proposed algorithms, maximizing the system resources utilization rate, etc. Further, as done in [75], convergent analysis, an alternative to competitive analysis, can also be used to evaluate the performance of on-line algorithms when the read-write request patterns are generally regular. Essentially, the choice of a performance metric and the method of performance evaluation usually stem from the underlying application requirements. We are interested to design object management algorithms and use various analyzing techniques to study the performance metrics mentioned above in the future. 112 Bibliography [1] B.Awerbuch, Y.Bartal, and A.Fiat, “Competitive Distributed File Allocation”, Proc. of the 25th Ann. ACM Symp on Theory of Computing, pp. 164-173, May 1993. [2] S.Albers, H.Koga, “Page Migration with Limited Local Memory Capacity”, Workshop on Algorithms and Data Structures, 1995. [3] H. Attiya and J. Welch, “Distributed Computing: Fundamentals, Simulations, and Advanced Topics”, McGraw-Hill Publishing, UK, 1998. [4] B.R.Badrinath, A.Acharya, and T.Imilelinski, “Designing Distributed Algorithms for Mobile Computing Networks”, Computers and Communications, 1994. [5] Y.Bartal, M.Charikar, and P.Indyk, “On Page Migration and Other Relaxed Task Systems”, In Proc. 8th Ann. ACM-SIAM Symp. on Discrete Algorithms, 1997. [6] Y.Bartal, A.Fiat, and Y.Rabani, “Competitive Algorithms for Distributed Data Management”, In Proc. of the 24th Ann. ACM Symp. on Theory of Computing, May 1992. [7] P. Bernstein, V. Hadzilacos, and N. Goodman, “Concurrency Control and Recovery in Database Systems”, Addison Wesley, Massachusetts, 1987. [8] Naser S.Barghouti and Gail E.Kaiser, “Concurrency Control in Advanced Database Applications”, ACM Computing Surveys, 23(3), pp.269-317, September 1991. Bibliography 113 [9] G.Barish and K.Obraczke, “World Wide Web Caching: Trends and Techniques”,IEEE Communications Magazine, Vol.38 5, pp. 178-184, May 2000. [10] S. Albers, “Competitive Online Algorithms”, BRICS LS-96-2,LSSN 1395-2048, September 1996. [11] D.L. Black and D.D.Sleator, “Competitive Algorithms for Replication and Migration Problems”, Technical Report CMU-CS-89-201, Department of Computer Science, Carnegie Mellon University, 1989. [12] V. Bharadwaj and E.M.Yew, “Network Caching Strategies for Reservation-based Multimedia Services on High-speed Networks”, Data&Knowledge Engineering 41, pp.85-103, 2002. [13] Beck Zaratian, “Microsoft Visual C++ Programmer’s Guide”, Microsoft Press, 1998. [14] George F.Coulouris, Jean Dollimore and Tim Kindberg, “Distributed Systems: Concepts and Design”, Addison-Wesley, 2001 [15] W.Cellary, E.Gelenbe, and T.Morzy, “Concurrency Control in Distributed Database Systems”, Elsevier Science Publishers, 1988. [16] T.H. Cormen, C.E. Leiserson, and R.L. Rivest, “Introduction to Algorithms”, The MIT Press, Cambridge USA, 1994. [17] M.Chrobak, L.Larmore, N.Reingold, and J.Westbrook, “Optimal Multiprocessor Migration Algorithms Using Work Functions”, in Proc. of the 4th Syrup. on Algorithms and Computation, 1993. [18] L.W.Dowdy and D.V.Foster, “Comparative Models of the File Assignment Problem”, ACM Computing Surveys, Vol.14, No.2, pp.287-313, June 1982. Bibliography 114 [19] Daniel Barbara, Hector Garcia-Molina, “Replicated Data Management in Mobile Environments: Anything New Under the Sun?”, on Application in Parallel and Distributed Computing, 1994. [20] B.David and G.Jane, “Distributed Database Systems”, Addison-Wesley Publishing Company, Inc., 1992. [21] M.H.Eich and S.H.Garad, “The Performance of Flow Graph Locking”, IEEE Transactions on Software Engineering”, Vol.16, No.4, pp.477-483, April 1990. [22] D.Gifford, “Weighted Voting for Replicated Data”, Proc. Seventh ACM Symp. Operation System Principles, pp. 150-162, 1979. [23] J.R.Haritsa, M.J.Carey, and M.Livny, “Earliest Deadline Scheduling for Real-Time Database Systems”, Proceedings of the Real-Time Systems Symposium, Dec. 1991. [24] A.Hac and T.Johnson, “A Study of Dynamic Load Balancing in a Distributed System”, Proceedings of the ACM SIGCOMM conference on Communications architecture & protocols, September 1986. [25] A.Hac and X.Jin, “Dynamic Load Balancing in a Distributed System Using a SenderInitiated Algorithm”, Proceedings of the 13th Conference on Local Computer Networks, pp.172-180, 1988. [26] Mounir Hamdi and Chi-Kin Lee, “Dynamic Load Balancing of Data Parallel Applications on a Distributed Network”, Proceedings of the 9th International Conference on Supercomputing, pp.170-179, 1995. [27] T. Hegazy and B. Ravindran, “Using Application Benefit for Proactive Resource Allocation in Asynchronous Real-Time Distributed systems”, IEEE Transactions on Computers, Vol.51, Num.8, pp.945-962, August 2002. Bibliography 115 [28] C. Hammacher, Z. Vranesic, and S. Zaky, “Computer Organization”, McGrawHill 5th edition, 2002. [29] K. Hwang and Z. Xu,“Scalable Parallel Computing-Technology, Architecture, Programming”, McGraw-Hill Publishers, 2000. [30] T. Imielinski and B. R. Badrinath. “Querying in Highly Mobile and Distributed Environment”, Proceeding of the 18th International Conference on Very Large Data Bases, pp.41-52, August 1992. [31] Gregory F. Johnson and Ambuj K. Singh, “Stable and Fault-Tolerant Object Allocation”, Proceedings of the 19th Annual ACM Symposium on Principles of Distributed Computing, pp.259-268, July 2000. [32] V.Kumar (Ed.), “Performance of Concurrency Control Mechanisms in Centralized Database Systems”, Prentice Hall, 1996. [33] K.Kalpakis, K.Dasgupta, and O.Wolfson, “Optimal Placement of Replicas in Trees with Read, Write, and Storage Costs”, IEEE Transactions on Parallel and Distributed Systems, Vol. 12, No. 6, pp.628-637, June 2001. [34] K.Hwang, “Advanced Computer Architecture: Parallelism, Scalability, Programmability”, McGraw-Hill Companies, Inc., 1993. [35] A. Karmouch,, E. Horlait, “Agent-based Multimedia Communication on Internet”, IEEE Canadian Conference on Electrical and Computer Engineering, Vol. 2, pp.734-737, 1998. [36] N.Katoh, T.Ibaraki, and T.Kameda, “Cautious Transaction Schedulers with Admission Control”, ACM Transactions on Database Systems”, Vol.10, No.2, pp.205-229, June 1985. [37] K.P.Eswaran, “Placement of records in a file and file allocation in a computer network”, IFIP Conference Information Proceedings, pp.304-307, August 1974. Bibliography 116 [38] T.Loukopoulos and I.Ahmad, “Static and Adaptive Data Replication Algorithms for Fast Information Access in Large Distributed Systems”, In proc. of the 20th IEEE International Conference on Distributed Computing Systems, pp.385-392, April 2000. [39] K.D.Levin and H.L.Morgan, “Optimizing Distributed Databases: A Framework for Research”, In Proceedings of AFIPS National Computer Conference, pp. 473-478, 1975. [40] C.Lund, N.Reingold, J.Westbrook, and D.Yan, “Competitive On-line Algorithms for Distributed Data Management”, SIAM Journal on Computing, vol.38, no.3, pp.1086-1111, Mar 1999. [41] Thomas D.C. Little and Dinesh Venkatesh, “Prospects for Interactive Video-OnDemand”, IEEE Multimedia, pp.14C24, 1994. [42] A.Leff, J.Wolf, and P.S.Yu, “LRU-based Replication Strategies in a LAN Remote Caching Architecture”, in Proc. 17th Annu. Conf. Local Comput. Networks, Minneapolis, MN, September 1992. [43] A.Leff, J.Wolf, and P.S.Yu, “Efficient LRU-based Buffering in a LAN Remote Caching Architecture”, IEEE Trans. Parallel and Distributed Systems, Vol.7, No.2, pp.191-206, Feb.1996. [44] J.C.Mogul, “Squeezing More Bits out of HTTP Caches”, IEEE Network, Vol.14 3, pp.6 -14, May-June 2000. [45] P.M.Melliar-Smith, L.E.Moser, V.Kalogeraki, and P.Narasimhan, “Realize: Resource Management for Soft Real-Time Distributed Systems”, Proceedings of the IEEE DARPA Information Survivability Conference and Exposition, pp.281 -293, January 2000. [46] M.S.Manasse, L.A.McGeoch, and D.D.Sleator, “Competitive Algorithms for On-line Problems”, Proceedings of the 20th Annual ACM Symposium on Theory of Computing, ACM, New York, pp.322-333, May 1988. Bibliography 117 [47] http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/HTML/ vccorehm.asp. [48] E.J.O’Neil, P.E.O’Neil, G.Weikum, “The LRU-K Page Replacement Algorithm for Database Disk Buffering”, Proceedings of ACM SIGMOD, 1993. [49] M.Tamer Ozsu, U.Dayal, and P.Valduriez, “Distributed Object Management”, Morgan Kaufmann Publishers, San Mateo, California, 1994. [50] Object Management Group, “The Common Object Request Broker: Architecture and Specification”, Revision 2.0, July 1995. [51] P.E.O’Neil, K.Ramamritham and C.Pu, “A Two-Phase Approach to Predictably Scheduling Real-Time Transactions”, Performance of Concurrency Control Algorithms in Centralized Database Systems, V.Kumar(ed.), Prentice-Hall, Englewood Cliffs, NJ, pp.494522, 1995. [52] T.M.Ozsu and P.Valduriez, “Principles of Distributed Database Systems”, Patrick Valduriez Prentice Hall, 1999. [53] P.Peddi and L.C.DiPippo, “A Replication Strategy for Distributed Real-Time ObjectOriented Databases”, Proceedings of the 5th IEEE International Symposium on ObjectOriented Real-Time Distributed Computing, pp.129-136, 2002. [54] Paul J.Fortier, “Database Systems Handbook”, McGraw-Hill Companies, Inc., 1997. [55] E.Pitoura and G.Samaras, “Data Management for Mobile Computing”, Kluwer Academic Publishers, 1998. [56] P.K.Reddy and S.Bhalla, “Asynchronous Operations in Distributed Concurrency Control”, IEEE Transactions on Knowledge and Data Engineering, Vol.15, No.3, pp.721-733, May 2003. Bibliography 118 [57] D.P.Reed, “Implementing Atomic Actions on Decentralized Data”, ACM Trans. on Computer Systems, Vol.1, No.1, pp.3-23, 1983. [58] K.Ranganathan, A.Iamnitchi, and I.Foster, “Improving Data Availability Through Dynamic Model-Driven Replication in Large Peer-to-Peer Communities”, In Global and Peerto-Peer Computing on Large Scale Distributed Systems Workshop, May 2002. [59] Y.Smaragdakis, S.Kaplan, and P.Wilson, “EELRU: Simple and Effective Adaptive Page Replacement”, Measurement and Modeling of Computer Systems, 1999. [60] Swarup Acharya and Stanley B.Zdonik, “An Efficient Scheme for Dynamic Data Replication”, Technical Report CS-93-43, Department of Computer Science, Brown University, September 1993. [61] D.D.Sleator and R.E.Tarjan, “Amortized Efficiency of List Update and Paging Rules”, Communications of the ACM, 28(2), pp.202-208, 1985. [62] A.Prasad Sistla,O.Wolfson and Y.Huang, “Minimization of Communication Cost Through Caching in Mobile Environments”, IEEE Transactions on Parallel and Distributed Systems, Vol. 9, No. 4, pp.378-390, April 1998. [63] P.Scheuermann, G.Weikum, P.Zabback, “Data Partitioning and Load Balancing in Parallel Disk Systems”, The Very Large Data Bases (VLDB) Journal, 7(1), pp.48-66, February 1998. [64] R.H.Thomas, “A Majority Consensus Approach to Concurrency Control for Multiple Copy Databases”, ACM Transactions on Database Systems, 4(2), pp.180-209, 1979. [65] R.Tewari, M.Dahlin, H.M.Vin, and J.Kay, “Beyond Hierarchies: Design Considerations for Distributed Caching on the Internet”, Technical Report CS98-04, Department of Computer Sciences, UT Austin, Texas, USA, May 1998. Bibliography 119 [66] Andrew S. Tanenbaum and Maarten Van Steen, “Distributed Systems: Principles and Paradigms”, Prentice Hall, 2002. [67] Mitchell D. Theys, Min Tan, Noah B. Beck, Howard Jay Siegel, and Michael Jurczyk, “A Mathematical Model and Scheduling Heuristics for Satisfying Prioritized Data Requests in an Oversubscribed Communication Network”, IEEE Transactions on Parallel and Distributed Systems, Vol. 11, No. 9, pp.969-988, September 2000. [68] B. Veeravalli, “Document Caching Policies on High-Speed Distributed Networks for Personalized Multimedia Services”, Proceedings of the 8th IEEE International Conference on Networks, pp. 215-219, September 2000. [69] P.Verissimo and L.Rodrigues, “Distributed Systems for System Architects”, Kluwer Academic Publishers, 2001. [70] J.Westbrook, “Randomized Algorithms for Multiprocessor Page Migration”, SIAM Journal on Computing, 1994. [71] J.Wang, “A Survey of Web Caching Schemes for the Internet”, ACM Computer Communication Review, Vol. 29, pp.36-46, October 1999. [72] Y. Wan and D. Du, “Video File Allocation over Disk Arrays for Video-On-Demand”, In Proceedings of International Conference for Multimedia Computing and Systems, 1996. [73] O. Wolfson and Y. Huang, “Competitive Analysis of Caching in Distributed Databases”, IEEE Transactions on Parallel and Distributed Systems, Vol.9, No.4, pp.391-409, April 1998. [74] Jun Wei, Tao He, and Tao Huang, “Challenges of Communication in Mobile Computing”, Proceedings of the Technology of Object-Oriented Languages and Systems, pp.196-203, 1998. Bibliography 120 [75] O.Wolfson and S.Jajodia, “Distributed Algorithms for Dynamic Replication of Data”, 11th ACM Symposium on Principles of Database Systems, pp.149-163, June 1992. [76] O.Wolfson and S. Jajodia, “An Algorithm for Dynamic Data Distribution”, Proc. Second Workshop Management of Replicated Data(WMRD-II), pp.62-65, 1992. [77] O. Wolfson, Shushil Jajodia, and Yixiu Huang, “An Adaptive Data Replication Algorithm”, ACM Transactions on Database Systems, 22(2), pp.255-314, 1997. [78] O.Wolfson and A.Milo, “The Multicast Policy and Its Relationship to Replicated Data Placement”, ACM Transactions on Database Systems, vol.16, no.1, pp.181-205, 1991. [79] Y. Won and J. Srivastava, “Strategic Replication of Video Files in a Distributed Environment”, International Journal of Multimedia Tools and Applications, Vol. 8, No.1, pp. 249-283, March 1999. [80] H.Xu, T.Furukawa, and Y.Shi, “Concurrency Control Based on Order Constraints in Advanced Database Systems”, In Proc. of International Symposium on Cooperative Database Systems for Advanced Applications (CODAS), pp.380-385, 1997. [81] Weining Zhang, Ke Wang, and Siu-Cheung Chau, “Data partition: A Practical Parallel Evaluation of Datalog Programs”, Proceedings of the First International Conference on Parallel and Distributed Information Systems, pp.98-105, 1991. 121 Author’s Publications [1] [Research Monograph] Wujuan Lin and Bharadwaj Veeravalli, “Object Management in Distributed Database Systems for Stationary and Mobile Computing Environments: A Competitive Approach”, Network Theory and Applications (NETA) Series, Vol. 12, Kluwer Academic publishers, Boston, USA, 2003. [2] Lin Wujuan and Bharadwaj Veeravalli, “An Adaptive Object Allocation and Replication Algorithm in Distributed Databases”, In the Proceedings of the 23rd IEEE International Conference on Distributed Computing Systems Workshops (ICDCSW’03), pp. 132-137, May 2003. [3] Lin Wujuan and Bharadwaj Veeravalli, “An Adaptive Object Management Algorithm for Distributed Databases in Mobile Computing Environments”, In the Proceedings of IEEE Joint Conference of the Fourth International Conference on Information, Communications & Signal Processing and Fourth Pacific-Rim Conference on Multimedia (ICICS-PCM 2003), Singapore, December 2003. [4] Lin Wujuan and Bharadwaj Veeravalli, “A Dynamic Object Allocation and Replication Algorithm for Distributed Databases with Centralized Control”, Submitted to The International Journal of Computers and Applications (IJCA), January 2003. [5] Lin Wujuan and Bharadwaj Veeravalli, “An Object Allocation and Replication Algorithm for Real-time Distributed Databases”, Submitted to the International Journal of Author’s Publications 122 Distributed and Parallel Databases, Kluwer Academic Publishers, June 2003. [6] Lin Wujuan and Bharadwaj Veeravalli, “An Adaptive Dynamic Algorithm for Data Allocation and Replication in Distributed Databases”, Submitted to IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans, August 2003. [7] Lin Wujuan and Bharadwaj Veeravalli, “Design and Analysis of a Dynamic Object Allocation and Replication Algorithm for Distributed Databases in Mobile Computing Environments”, Submitted to ACM Transactions on Database Systems (TODS), November 2003. [...]... in the rest of this thesis were presented In the next chapter, we first consider the object allocation and replication issues in SCEs 16 Chapter 3 Object Management in Stationary Computing Environments In this chapter, we address the object allocation and replication issues of OMP in DDBSs in the application domain of SCEs Designing an intelligent and efficient object allocation and replication on-line... Finally, we carried out experiments to study the performance of ADRW algorithm under several in uencing conditions in a SCE We conducted detailed performance analysis and comparisons in the experiments The experimental results give more insights on designing object allocation and replication strategies for DDBSs In conclusion, our research contribution lies in designing adaptive object allocation and. .. issues of an OMP In other words, an on-line object allocation and replication algorithm recommends a set of processors, often referred to as an object allocation scheme, that need to have copies of an object 1.2 Issues to Be Studied and Main Contributions The issues mentioned in Section 1.1 considerably motivate us to design cost-effective algorithms for object allocation and replication issues in DDBSs In. .. system performance In this thesis, we concentrate on exposing the underlying key challenges in designing on-line algorithms to handle unpredictable requests that arrive at a DDBS We design several dynamic online algorithms for the object allocation and object replication issues which form a part of the OMP Our objective is to provide a theoretical framework and rigorously analyze the performance of the... network Thus, in such a scenario, one of the main problems is in designing efficient policies to handle on-line requests arriving at the system with a minimum cost and maintain the consistency of multiple replicas of objects in various locations in the network As mentioned above, replication increases the object availability by allowing many nodes to service several requests for the same object concurrently... performance of the algorithms, it is sufficient to consider a single object and analyze the behavior of the algorithms under several in uencing factors Finally, we will simply discuss on the system reliability issue for ADRW algorithm in terms of the failure and recovery 3.1 Preliminaries and Problem Formulation As far as object allocation and replication issues are concerned, both static and dynamic algorithms. .. object allocation scheme in Section 1.1 In fact, an OMP for a DDBS attempts to modify or use this allocation scheme information to seek the most recent copy of an object [14, 49, 52, 69] The object allocation scheme can be a dynamic quantity depending on the strategy used in the design of OMP By and large, most of the object allocation and replication strategies are geared towards efficient ways of managing... application domains, these two issues may obtain different concerns and pose various challenges to the algorithm/system designers We consider following two distinct application domains in this study, i.e., DDBSs in Stationary Computing Environments (SCEs) and Mobile Computing Environments (MCEs) Traditionally, a DDBS in a SCE consists of several stationary nodes in the system The location of a node in the system... be distributed/ managed in several locations in the system These issues include, • Object Allocation: Determining the locations to hold an object when the object is created (Choosing vantage locations for the respective objects) • Object Location: Determining the locations of an object whenever an end user wishes to access it (Equivalent to searching locations to find the desired objects) • Object Replication: ... their corresponding invalidatelists except processor pi (if pi is in an invalidate-list) to invalidate the outdated replicas of object o It may be noted that if processor pi ∈ S(o) and pi is not in any invalidate-list for object o, then pi will indicate some server in S(o) to add itself into the corresponding invalidate-list The purpose of this process is to invalidate the copy of object o in pi when DA . DESIGN AND ANALYSIS OF OBJECT ALLOCATION AND REPLICATION ALGORITHMS IN DISTRIBUTED DATABASES FOR STATIONARY AND MOBILE COMPUTING SYSTEMS LIN WUJUAN (B.Eng., Xi’an Jiaotong. insights on designing object allocation and replication strategies for DDBSs. In conclusion, our research contribution lies in designing adaptive object allocation and repli- cation algorithms and evaluating. one of the main problems is in designing efficient policies to handle on-line requests arriving at the system with a minimum cost and maintain the consistency of multiple replicas of objects in