Adaptable pointer swizzling strategies in object bases: design,realization, and quantitative analysis

In this article, different techniques for pointer swizzling are classified and evaluated for optimizing the access to main-memory resident persistent objects. To speed up the access along inter-object references, the persistent pointers in the form of unique object identifiers (OIDs) are transformed (swizzled) into main-memory pointers (addresses). Pointer swizzling techniques can be divided into two classes: (1) those that allow replacement of swizzled objects from the buffer before the end of an application program, and (2) those that rule out the displacement of swizzled objects. The first class (i.e., techniques that take precautions for the replacement of swizzled objects) has not yet been thoroughly investigated. Four different pointer swizzling techniques allowing object replacement are investigated and compared with the performance of an object manager employing no pointer swizzling. The extensive qualitative and quantitative evaluation—only part of which could be presented in this article—demonstrate that there is noone superior pointer swizzling strategy forall application profiles. Therefore, an adaptable object base run-time system is devised that employs the full range of pointer swizzling strategies, depending on the application profile characteristics that are determined by, for example, monitoring in combination with sampling, user specifications, and/or program analysis.     

基于虚存映射的对象存储系统的设计与实现

多媒体智能数据库系统ＭＩＤＳ（ＭｕｌｔｉｍｅｄｉａＩｎｔｅｌｌｉｇｅｎｔＤａｔａｂａｓｅＳｙｓｔｅｍ）的存储系统是一个强大的，高性能的对象存储系统，该存储系统是ＭＩＤＳ引擎的核心，它是基于虚存映身的存储系统，在Ｃｌｉｅｎｔ／Ｓｅｒｖｅｒ的体系结构下，为数据库提供了大容量，复杂的，共享的，多革体数据存取，然而又不失面的对象数据模型的种种优点，在本文中，我们将介绍基于虚存映射的对象存储系统的基本原理以     

Object-oriented query language access to relational databases: A semantic framework for query translation

This research investigates and approach to query processing in a multidatabase system that uses an objectoriented model to capture the semantics of other data models. The object-oriented model is used to construct a global schema, defining an integrated view of the different schemas in the environment. The model is also used as a self-describing model to build a meta-database for storing information about the global schema. A unique aspect of this work is that the object-oriented model is used to describe the different data models of the multidatabase environment, thereby extending the meta database with semantic information about the local schemas. With the global and local schemas all represented in an object-oriented form, structural mappings between the global schema and each local schema are then easily supported. An object algebra then provides a query language for expressing global queries, using the structural mappings to translate object algebra queries into SQL queries over local relational schema. The advantage of using an object algebra is that the object-oriented database can be viewed as a blackboard for temporary storage of local data and for establishing relationships between different databases. The object algebra can be used to directly retrieve temporarily-stored data from the object-oriented database or to transparently retrieve data from local sources using the translation process described in this paper.     

A formal approach to object-oriented databases

Object-oriented database systems are the focus of current research and development efforts. Yet, there is no commonly accepted object model, nor is it clear whether such a model can be developed. This paper reports on efforts to develop a formal framework that contains most features found in current object oriented database systems. The framework contains two parts. The first is a structural object model, including concepts such as structured objects, identity, and some form of inheritance. For this model, we explain the distinction between values and (abstract) objects, describe a system as a directed graph, and discuss declarative languages. The second part deals with higher-order concepts, such as classes and functions as data, methods, and inheritance. This part is a sketch, and leaves many issues unresolved. Throughout the paper, the emphasis is on logic-oriented modeling.     

C＋＋对象的持久化中的问题和解决方案

多媒体智能数据库系统ＭＩＤＳ（ｍｕｌｔｉｍｅｄｉａｉｎｔｅｌｌｉｇｅｎｔｄａｔａｂａｓｅｓｙｓｔｅｍ）是一个对象数据库管理系统．它的数据库编程语言是Ｐ＋＋，Ｐ＋＋个语言是一种基于Ｃ＋＋的语言．作者在Ｐ＋＋的实现过程中遇到了以下问题：首先，Ｃ＋＋的指针有二义性，它无法在语义上区分成员指针和引用指针，以及易失性指针和持久性指针，从而给事务管理中的对象加锁及其它方面带来问题．其次，具有虚拟函数或虚拟基类的对象中含有指向内存中的指针，而这些指针不是由程序员定义的．如果Ｃ＋个对象被持久化，这些指针在不同的程序调用中有可能无效．最后，如果作者用文件系统调用来存储对象，那么必须要设计复杂的Ｃａｃｈｅ系统和做大量对象的格式转化工作，这需要大量的空间和时间，所以他们采用了另外的一种方法──基于虚拟内存空间映射的存储方案．     

映射对象到关系数据库

面向对象技术是当今最流行的软件开发方法,关系模型是最常用的一种数据模型。如何用关系数据库实现持久对象的存储成为信息系统成功的关键。本文对映射对象模型到存储数据的关系模型的方法进行了探讨。     

Using object deputy model to prepare data for data warehousing

Providing integrated access to multiple, distributed, heterogeneous databases and other information sources has become one of the leading issues in database research and the industry. One of the most effective approaches is to extract and integrate information of interest from each source in advance and store them in a centralized repository (known as a data warehouse). When a query is posed, it is evaluated directly at the warehouse without accessing the original information sources. One of the techniques that this approach uses to improve the efficiency of query processing is materialized view(s). Essentially, materialized views are used for data warehouses, and various methods for relational databases have been developed. In this paper, we first discuss an object deputy approach to realize materialized object views for data warehouses which can also incorporate object-oriented databases. A framework has been developed using Smalltalk to prepare data for data warehousing, in which an object deputy model and database connecting tools have been implemented. The object deputy model can provide an easy-to-use way to resolve inconsistency and conflicts while preparing data for data warehousing, as evidenced by our empirical study.     

基于．NET平台ORM技术阳光采购系统的设计

在基于．NET平台的管理信息系统中,面向对象的业务处理和关系数据存储存在着一定的矛盾。在分析．NET平台的数据库处理的技术基础上,设计了一个对象／关系数据库映射模式,并结合这个模式创建了一个通用的数据操作框架CFC,实现了对象／关系数据库映射。CFC框架使软件开发人员能够以面向对象的方法进行对象数据存取,而不必了解后台关系数据库的实现。该框架实际应用到了广东电网公司阳光采购系统。     

基于内存映射文件的复杂对象快速读取方法

当前内存数据库(NoSQL)、嵌入式数据库技术在高并发高性能系统中得到了广泛的应用,但对于复杂对象数据的读取效率仍然低下,研究发现主要性能瓶颈有两个:一是内核态与用户态间的内存拷贝,拷贝消耗时间与复杂对象的数据量成线性增长;二是从数据库数据到运行时数据对象的格式转化操作,不但需要开辟新的内存空间存储运行时数据对象,而且还需要解析原始数据并拷贝至新对象之中。为此,提出了一种基于内存映射(memory mapping)文件的复杂对象共享读取方法。借助内存映射文件与自定义内存分配器,实现了结构复杂的C++标准模板库容器对象跨进程无拷贝、无格式转化的共享,有效降低了数据读取延时。通过性能的分析比较表明,与NoSQL内存数据库、嵌入式数据库比,读取性能效率提升10倍以上。再加上底层技术成熟稳定,复用了标准模板库,具有开发成本低、可维护性强、实用性高等优点,因此,适用于高并发高性能的高可用后台服务系统。     

对象模式到关系模式转换的软件框架研究

数据在信息系统中的使用面临着这样一个困难，即数据存储是以关系模型为基础，而软件开发以对象模型来进行，造成了软件开发中数据访问技术的不和谐。在软件实现上陷入两种模式的转换工作，破坏面向对象语言的面向对象性，造成开发效率低下，代码重用率变低。提出并建立了一个软件框架，利用它提供的处于关系数据库和客户端之间的API来进行基于对象的数据库访问，充分发挥两种不同模式的优点，以提高软件开发的效率。     

面向CFD应用的Intel持久内存性能评估

在科学计算领域,数据规模随着数值模拟精度要求的提高而快速增长,以DRAM为主存的传统方案由于成本高而难以扩展容量,近年来越来越被关注的持久内存技术有望解决这一问题。持久内存是在DRAM和SSD之间的补充,相比DRAM,持久内存具有容量大、性价比高的优点,但是性能也相对较低。为测试持久内存的应用性能,面向科学计算的一个重要领域——计算流体力学(CFD),对Intel持久内存进行性能评估。实验中,持久内存采用了最易于使用的内存模式,源码不需要任何修改,测试程序涵盖内存基准测试和3种常见的CFD算法,实验结果表明,在内存模式下,对不同CFD算法,相比纯DRAM的配置,持久内存的引入会带来一定的性能损失,且该损失随数据规模的增加而增大;另一方面,持久内存的部署使单服务器能支撑超大数据规模的数值模拟。     

地理信息系统与数据库结合研究

鉴于地理信息系统需要对外提供灵活高效的地理信息访问途径,因此采用数据库管理地理信息是发展趋势,目前存在关系型,对象-关系型,面向对象型3类数据库产品,在讨论它们与地理信息系统结合可行性的基础上,重点介绍了纯关系模型,空间扩展模型,包装模型3种结合方案,还通过对这3种方案实现原理的分析及其优缺点的比较,指出了各方面的应用前景,并由此归纳出,在GIS与数据库结合中需要注意的问题及解决这些问题的关键技术。     

An object-based architecture for biomedical expert database systems

Objects play a major role in both database and artificial intelligence research. In this paper, we present a novel architecture for expert database systems that introduces an object-based interface between relational databases and expert systems. We exploit a semantic model of the database structure to map relations automatically into object templates, where each template can be a complex combination of join and projection operations. Moreover, we arrange the templates into object networks that represent different views of the same database. Separate processes instantiate those templates using data from the base relations, cache the resulting instances in main memory, navigate through a given network's objects, and update the database according to changes made at the object layer. In the context of an immunologic-research application, we demonstrate the capabilities of a prototype implementation of the architecture. The resulting model provides enhanced tools for database structuring and manipulation. In addition, this architecture supports efficient bidirectional communication between database and expert systems through the shared object layer.     

A persistent complex object database language

To support a number of emerging applications we propose a persistent complex object database language which is “value” based. The object space is built through the set and tuple object constructors, with practically no normalization constraints on the nesting levels. The proposed persistent database language can serve both as an interface language to the database management system in itself or as the target language of a conceptual data language. Some of the features of the language include: (i) Notion of “path's” capturing logical references to retrieve and update the persistent database, (ii) a rich set of allowable schema types, including disjuncts, and (iii) a collection of powerful operator constructors for sets.     

The application of data types to database semantic integrity

Data type concepts are used to investigate the extent to which database semantic integrity can be defined and ensured through database structures. Database and datatype concepts are extended mutually to improve the semantic capabilities of both database models and data type systems and to resolve apparent discrepencies between databases and programming languages. To meet database needs, data structuring is developed to form an algebra of data types. A semantically rich database model is used to show that database models can be expressed in terms of data types. Finally, a schema specification language is presented to demonstrate the power of data type tools for the definition of database schemas and for the maintenance of database semantic integrity.     

基于持久性内存的属性图存储系统

属性图是一种流行的图数据模型, 在各种图系统中得到了广泛应用. 然而, 面向事务型负载的图数据库系统在执行图分析任务的场景下面临着高延迟等挑战. 传统的图分析系统往往是基于简单图模型, 而且大多不支持图的事务型负载. 因此, 迫切需要一个能够在属性图上高效处理事务型负载和图分析任务的图存储系统. 持久性内存的问世, 使得我们有机会重新设计图存储系统, 以充分发挥这种设备的特点. 为此, 本文提出了一种基于持久性内存的属性图存储系统, 名为TAG. TAG采用了一种新颖的混合架构的图存储方式, 以充分发挥持久性内存和主存的优势. 其次, 通过拓扑和索引结合的方式, 将图的拓扑嵌入到系统的索引中以加速图的拓扑查询. 最后, TAG通过基于标签的方式来组织图的属性数据, 进一步优化图的属性访问. 实验结果表明, TAG显著优于其他图数据库系统, 与图分析系统相比, TAG也有着相近的性能表现.     

Measuring and modelling the performance of a parallel ODMG compliant object database server

Object database management systems (ODBMSs) are now established as the database management technology of choice for a range of challenging data intensive applications. Furthermore, the applications associated with object databases typically have stringent performance requirements, and some are associated with very large data sets. An important feature for the performance of object databases is the speed at which relationships can be explored. In queries, this depends on the effectiveness of different join algorithms into which queries that follow relationships can be compiled. This paper presents a performance evaluation of the Polar parallel object database system, focusing in particular on the performance of parallel join algorithms. Polar is a parallel, shared‐nothing implementation of the Object Database Management Group (ODMG) standard for object databases. The paper presents an empirical evaluation of queries expressed in the ODMG Query Language (OQL), as well as a cost model for the parallel algebra that is used to evaluate OQL queries. The cost model is validated against the empirical results for a collection of queries using four different join algorithms, one that is value based and three that are pointer based. Copyright © 2005 John Wiley & Sons, Ltd.     

Integration of CORBA and object relational databases

CORBA is widely accepted as the open international standard for modelling and building comprehensive distributed systems. In most cases, CORBA architects have adopted relational databases for storage of persistent data. Among the issues that usually face architecture designers considering how to combine CORBA and standard relational database standards are fault tolerance, performance, and the extensibility and scalability of the systems. The research team involved with this paper found that the ODMG object database concept is useful to solve the issues encountered when integrating CORBA and relational database standards. The reference architecture, which the team devises, integrates CORBA and relational databases without compromise on the necessary transactional properties. The CORBA standard object transaction service and concurrency control service are reused. The team also develop an object relational data modelling tool—Latte—that supports the overall design intention as well the development paradigms for the proposed architecture. The implementation of the system is useful to CORBA, ODMG, and relational database architects because it provides a unified modelling and programming paradigm capable of solving the problems of managing mission-critical distributed data. Thus, we present a case study of combining different international standards to build a comprehensive system.     

Trigger inheritance and overriding in an active object databasesystem

An active database is a database in which some operations are automatically executed when specified events happen and particular conditions are met. Several systems supporting active rules in an object oriented data model have been proposed. However, several issues related to the integration of triggers with object oriented modeling concepts have not been satisfactorily addressed. We discuss issues related to trigger inheritance and refinement in the context of the Chimera active object oriented data model. In particular, we introduce a semantics for an active object language that takes into account trigger inheritance and supports trigger overriding. Moreover, we state conditions on trigger overriding ensuring that trigger semantics is preserved in subclasses