T. Patrick Martin

Automatically classifying database workloads

The type of the workload on a database management system (DBMS) is a key consideration in tuning the system. Allocations for resources such as main memory can be very different depending on whether the workload type is Online Transaction Processing (OLTP) or Decision Support System (DSS). In this paper, we present an approach to automatically identifying a DBMS workload as either OLTP or DSS. We build a classification model based on the most significant workload characteristics that differentiate OLTP from DSS, and then use the model to identify any change in the workload type. We construct a workload classifier from the Browsing and Ordering profiles of the TPC-W benchmark. Experiments with an industry-supplied workload show that our classifier accurately identifies the mix of OLTP and DSS work within an application workload.

Today's DBMSs: how autonomic are they

The increasing complexity of database management systems (DBMSs) and the dearth of their experienced administrators make an urgent call for an autonomic DBMS that is capable of managing and maintaining itself. In this paper, we examine the characteristics that a DBMS should have in order to be considered autonomic and assess the position of todays commercial DBMSs such as DB2, SQL Server, and Oracle.

Data caching strategies for distributed full text retrieval systems

Abstract We discuss and evaluate several strategies for caching data in a distributed full text retrieval system. The evaluation of the strategies is carried out using a simulation model of a distributed full text retrieval system. We describe the development of this simulation model which was based on a distributed prototype of an existing full text retrieval system, namely Ful/Text [1]. The simulation model was used to analyze the effect data caching would have on the performance of such a system and to show the relationship between several properties of the workload and the effectiveness of data caching.

Strategies for building distributed information retrieval systems

Abstract In this article we discuss the need for distributed information retrieval Systems. A number of possible configurations are presented. A general approach to the design of such systems is discussed. A prototype implementation is described together with the experiences gained from this implementation.

Supporting Browsing of Large Knowledge Bases

This paper examines the problem of browsing in knowledgebased systems. Current systems provide inadequate support for browsing large knowledge bases. Three areas where this support must be enhanced are the user interface, the facilities for querying the knowledge base and the facilities for exploring the knowledge base. Techniques to provide this enhanced support are described. A prototype system for browsing large framebased knowledge bases is presented.1

Remote operations across a network of small computers

This paper discusses the design of a Remote Operation Call (ROC) mechanism. ROCs are a generalisation of the remote procedure call concept. They provide for a wider variety of remote calls, such as asynchronous, directed and multicast calls. An implementation of ROCs on a network of personal computers is also described.

Remote procedure call facility for a PC environment

Abstract Remote Procedure Calls are a common mechanism for remote execution. This paper proposes the use of RPCs to better integrate computers into distributed computing environments. Homogeneous RPCs, including procedure stubs and the RPC communication package, are discussed in detail. The implementation of a RPC facility for hetrogeneous environments consisting of a network of personal computers and Sun/Unix machines is also described, including aspects such as data conversion filters, session level interface and gateway configuration. Related work to this area is also outlined, and the success of integrating PCs into distributed computing environments is discussed.

A Knowledge-Based System for Fault Diagnosis in Real-Time Engineering Applications

We describe the development of a knowledge-based system for fault diagnosis in complex real-time engineering devices. The characteristics of this type of application require that the system combine reasoning, database, and real-time data acquisition capabilities. We concentrate on an explanation of the knowledge and database components of this system.

Abstraction hierarchies in top-down design

This paper describes a program design discipline that has successfully produced well-modularized programs. The basic approach is to apply, in a uniform way, the concepts of data and procedural abstraction in a top-down decomposition during the initial programming-in-the-large phase of construction. This combination of data and procedural abstraction, called hybrid abstraction, views the system as composed almost entirely of abstract objects. The resulting structural design is partially expressed as a syntactic specification for a set of modules (a module being a set of objects) having no directly shared global data. In addition to the syntactic specification, the design expresses the abstract type decomposition and the structural relationships of the modules. Three relations on the set of modules define the hierarchies is called by, implements, and obtains resources from. These relations define the discipline and characterize the class of designs possible. Modules in practice are used in two different ways: single-instance use and multiple-instance use. Abstract objects can be of two different kinds: collection objects and singular objects. The relationships of these variations in module construction are illustrated. Three moderately large examples are shown. euclid is used to illustrate the discipline.

Querying and Exploring Large Knowledge Bases

This paper presents three aspects of information retrieval from large knowledge bases: the abstraction mechanisms, the modes, and the interface model. To reduce retrieval complexity, we investigate the view and viewpoint abstraction mechanisms. We consider and discuss querying and exploring as two major modes of knowledge base retrieval.