Performance Analysis of Distributed Object Middleware Technologies

Each day new computers around the world connects to the Internet or some network. The increasing number of people and computers on the Internet has lead to a demand for more services in different domains that can be accessed from many locations in the network.When the computers communicate they use different kinds of protocols to be able to deliver a service. One of these protocol families are remote procedure calls between computers. Remote procedure calls has been around for quite some time but it is with the Internet that its usage has increased a lot and especially in its object oriented form which comes from the fact that object oriented programming has become a popular choice amongst programmers. When a programmer has to choose a distributed object middleware there is a lot to take into consideration and one of those things is performance.This master thesis aims to give a performance comparison between different distributed object middleware technologies and give an overview of the performance difference between them and make it easier for a programmer to choose one of the technologies when performance is an important factor. In this thesis we have evaluated the performance of CORBA, DCOM, RMI, RMI-IIOP, Remoting-TCP and Remoting-HTTP. The results we have seen from this evaluation is that DCOM and RMI…


PART 1 Introduction
1.1 Motivation for writing this thesis
1.2 Problem definition
1.3 Scope and Limitations
1.4 Methodology
1.5 Main Contribution
1.6 Thesis Outline
1.7 Definitions
PART 2 Background
2.1 Remote Procedure Calls
2.1.1 Overview
2.1.2 Passing Parameters
2.1.3 Parameter Representation
2.1.4 Object Oriented Approach
2.2 .NET Remoting
2.2.1 Motivation
2.2.2 Overvie
2.2.3 How .NET Remoting works
2.2.4 Remote Object Proxies
2.2.5 Easy Deployment of Distributed Objects
2.2.6 Remoting-TCP
2.2.7 Remoting-HTTP With Simple Object Access Protocol
2.3 Distributed Component Object Model
2.3.1 Motivation
2.3.2 Overview
2.3.3 How COM/DCOM Works
2.3.4 The Windows Registry
2.3.5 Service Control Manager (SCM)
2.3.6 Interface Description Language
2.3.7 Fallback Functionality
2.4 Common Object Request Broker Architecture
2.4.1 Motivation
2.4.2 Overview
2.4.3 How CORBA Works
2.4.4 Object Request Broker
2.4.5 Interface Description Language
2.5 Remote Method Invocation
2.5.1 Overview
2.5.2 Motivation
2.5.3 How RMI Works
2.5.4 Interfaces
2.5.5 The RMI Registry
2.5.6 RMI-IIOP
2.6 Summary
2.6.1 Architecture
2.6.2 Deployment
2.6.3 Run-time
PART 3 Experimental setup
3.1 Test environment
3.1.1 Hardware
3.1.2 Software
3.2 The Test Framework
3.2.1 Requirements of the Framework
3.2.2 High Resolution Timers in the Test Framework
3.2.3 The Design of the Framework
3.3 Test Data
3.3.1 Performance Tests
3.3.2 Overhead Tests
3.3.3 Data Type Mappings Between the Architectures
3.3.4 Motivation of choices in test data
PART 4 Results
4.1 Measuring Network Traffic
4.1.1 Motivation
4.1.2 Method
4.1.3 Simple/Primitive Data Types
4.1.4 Array Size Test for Primitive Data Types
4.1.5 String Size Test
4.1.6 Summary
4.2 Time based performance results
4.2.1 Motivation
4.2.2 Method
4.2.3 Primitives and Objects
4.2.4 Strings and Arrays
4.2.5 Summary of the throughput tests
PART 5 Discussion
5.0.1 Working with the Different Distributed Object Middleware:s
PART 6 Related Work
PART 7 Future Work
7.1 Other Distributed Object Middleware & RPC Middleware
7.1.1 Web Services
7.1.2 XML-RPC
PART 8 Conclusion
PART 9 References

Author: Per Arneng, Richard Bladh

Source: Blekinge Institute of Technology

Download URL 2: Visit Now

Leave a Comment