Table of Contents:

Definition and Overview

1. Introduction

2. Softswitch Architectural Overview

3. Network Element Failure Scenarios in the Network Office

4. High-Availability Techniques for the Network Office

5. Software Failure Analysis

6. Conclusions

Self-Test

Glossary

PDF of this tutorial

Softswitches, though critical to implementing the network office, are relatively new to the industry. This makes it difficult yet crucial to identify the most likely causes of loss of availability with a high degree of certainty. This analysis can be facilitated through the use of experience with other complex distributed systems. Some assumptions can be made:

• First, network outages are likely to be the leading cause of failures. This could be the result of wide-area network (WAN) communication link outages, malfunctioning router software, or simply the wrong cable disconnected at a patch panel somewhere in the network.
• Second, software problems will occur. The name softswitch itself implies the central role of software inhis architecture. Software problems that do arise most often manifest themselves when a new version of software is deployed in a live network. The quickest method of dealing with this failure scenario is to roll back to a previous version of software. High-availability goals can only be achieved by an architecture that supports multiple versions of software in the softswitch components, as well as the ability to rapidly change between these versions.
• Third, hardware reliability issues with individual components are infrequent sources of failure, even without fault-tolerant platforms. Sun workstations are the most common hardware platforms for softswitches, and their reliability easily exceeds the requirements of softswitches. Hardware failures that do occur can be solved with software redundancy across multiple hardware platforms.