Hardware trends have produced an increasing disparity between processor speeds and memory access times. While a variety of techniques for tolerating or reducing memory latency have been proposed, these are rarely successful for pointer-manipulating programs. This paper explores a complementary approach that attacks the source (poor reference locality) of the problem rather than its manifestation (memory latency). It demonstrates that careful data organization and layout provides an essential mechanism to improve the cache locality of pointer-manipulating programs and consequently, their performance. It explores two placement technique-lustering and colorinet improve cache performance by increasing a pointer structure’s spatial and temporal locality, and by reducing cache-conflicts. To reduce the cost of applying these techniques, this paper discusses two strategies-cache-conscious reorganization and cacheconscious allocation--and describes two semi-automatic toolsccmorph and ccmalloc-that use these strategies to produce cache-conscious pointer structure layouts. ccmorph is a transparent tree reorganizer that utilizes topology information to cluster and color the structure. ccmalloc is a cache-conscious heap allocator that attempts to co-locate contemporaneously accessed data elements in the same physical cache block. Our evaluations, with microbenchmarks, several small benchmarks, and a couple of large real-world applications, demonstrate that the cache-conscious structure layouts produced by ccmorph and ccmalloc offer large performance benefit-n most cases, significantly outperforming state-of-the-art prefetching.

Pointer swizzling is the conversion of database objects between an external form (object identifiers) and an internal form (direct memory pointers). Swizzling is used in some object-oriented databases, persistent object stores, and persistent and database programming language implementations to speed manipulation of memory resident data. Here we describe a simplifying model of application behavior, revealing those aspects where swizzling is most relevant in both benefits and costs. The model has a number of parameters, whichwehavemeasured for a particular instance of the Mneme persistent object store, varying the swizzling technique used. The results confirm most of the intuitive, qualitative tradeoffs, with the quantitative data showing that some performance differences between schemes are smaller than might be expected. However, there are some interesting effects that run counter to naive intuition, most of which we explain using deeper analysis of the algorithms and data struc...

The goal of this work is to integrate in a general framework the different query optimization techniques that have been proposed in the object-oriented context. As a first step, we focus essentially on the logical aspect of query optimization. In this paper, we propose a formalism (i) that unifies different rewriting formalisms, (ii) that allows easy and exhaustive factorization of duplicated subqueries, and (iii) that supports heuristics in order to reduce the optimization rewriting phase. 1 Introduction Many declarative query languages for object-oriented database management systems have been proposed in the last few years (e.g. [4, 5, 9]). Currently, a very serious concern is their optimization. Interesting techniques have been imported from other environments or developed for this context. However, although complementary, these techniques are often supported by distinct formalisms. The implementation of a query optimizer is thus still an awkward and delicate operation. Clearly, i...

We investigate the performance of some of the best-known object clustering algorithms on four different workloads based upon the Tektronix benchmark. For all four workloads, stochastic clustering gave the best performance for a variety of performance metrics. Since stochastic clustering is computationally expensive, it is interesting that for every workload there was at least one cheaper clustering algorithm that matched or almost matched stochastic clustering. Unfortunately, for each workload, the algorithm that approximated stochastic clustering was different. Our experiments also demonstrated that even when the workload and object graph are fixed, the choice of the clustering algorithm depends upon the goals of the system. For example, if the goal is to perform well on traversals of small portions of the database starting with a cold cache, the important metric is the per-traversal expansion factor, and a well-chosen placement tree will be nearly optimal; if the goal is to achieve a...

Thor is a new object-oriented database management system (OODBMS), intended to be used in heterogeneous distributed systems to allow programs written in different programming languages to share objects in a convenient manner. Thor objects are persistent in spite of failures, are highly likely to be accessible whenever they are needed, and can be structured to reflect the kinds of information of interest to users. Thor combines the advantages of the object-oriented approach with those of database systems: users can store and manipulate objects that capture the semantics of their applications, and can also access objects via queries. Thor is an ongoing project, and this paper is a snapshot: we describe our first design and a partial implementation of that design. This design is primarily concerned with issues related to the implementation of an OODBMS as a distributed system. 1 INTRODUCTION Distributed systems contain different kinds of computers, and users write programs in different...

Object-oriented databases store many small objects on disks. Disks perform poorly when reading and writing individual small objects. This thesis presents a new storage management architecture that substantially improves disk performance of a distributed object-oriented database system. The storage architecture is built around a large modified object buffer (MOB) that is stored in primary memory. The MOB provides volatile storage for modified objects. Modified objects are placed in the MOB instead of being immediately written out to disk. Modifications are written to disk lazily as the MOB fills up and space is required for new modifications. The MOB improves performance because even if an object is modified many times in a short period of time, the object has to be written out to disk only once. Furthermore, by the time an object modification has to be flushed from the MOB, many modifications to other objects on the same page may have accumulated. All of these modifications can be writ...

Many client-server object-oriented database systems (OODBs) run applications at clients and perform all accesses on cached copies of database objects. Moving both data and computation to the clients can improve response time, throughput, and scalability. For applications with good locality of reference, retaining cached state across transaction boundaries can result in further performance and scaling benefits. This thesis examines the question of what concurrency control scheme is best able to realize these potential benefits. It describes a new optimistic concurrency control scheme called AOCC (Adaptive Optimistic Concurrency Control) and compares its performance with that of ACBL (Adaptive-Granularity Callback Locking), the scheme shown to have the best performance in previous studies. Like all optimistic schemes, AOCC synchronizes transactions at the commit point, aborting transactions when synchronization fails; ACBL, like other locking schemes, synchronizes transactions while they execute. Earlier

In this thesis we address the problem of providing efficient processing of queries in the extensible environment induced by object-oriented databases. We define a framework for query processing in an object-oriented database and develop designs for major components of this framework. The framework encompasses an object-oriented data model, an algebra to query over that model, transformation rules for the algebra, an internal representation for queries expressed in the algebra, a cost model for analyzing query expressions, and an architecture for an extensible query optimizer. The major contributions of this thesis are an algebra and transformation rules, a representation, and an architecture for extensible query optimization. We show how these components fit into the framework and interact with each other. The EQUAL query algebra presented in this thesis is the first query algebra for object-oriented database systems to be completely consistent with data abstraction, and one of the few...

In an orthogonally persistent programming system the longevity of data is independent of its other attributes. The advantages of persistence may be seen primarily in the areas of data modelling and protection resulting from simpler semantics and reduced complexity. These have been verified by the first implementations of persistent languages, typically consisting of a persistent store, a run-time system and a compiler that produces programs that may access and manipulate the persistent environment. This thesis demonstrates that persistence can deliver many further benefits to the programming process when applied to software construction and execution. To support the thesis, a persistent environment has been extended with all the components necessary to support program construction and execution entirely within the persistent environment. This is the first known example of a strongly-typed integrated persistent programming environment. The keystone of this work is the construction of ...

We describe an eficient server-based algorithm for garbage collecting object-oriented databases in a client/server environment. The algorithm is incremen-tal and runs concurrently with client transactions. Un-like previous algorithms, it does not hold any locks on data and does not require callbacks to clients. It is fault tolerant, but performs very little logging. The algorithm has been designed to be integrated into existing OODB systems, and therefore it works with standard implemen-tation techniques such as two-phase locking and write-ahead-logging. In addition, it supports client-server per-formance optimizations such as client caching and flexible management of client buffers. We describe an implemen-tation of the algorithm in the EXODUS storage manager and present results from an initial performance study. 1