Chapter 13. What's New in the Tornado Overview System

Table of Contents
13.1. Introduction
13.2. Faster Design
13.3. Overview Semantics
13.4. I/O Optimization

13.1. Introduction

With the introduction of Tornado, the primary issue with the current overview indexing system is o files growing beyond the 2GB limit. Since Tornado Front End can represent the contents of many Tornado Back Ends, the indexing for all the articles was consuming the existing overview system.

Additionally, the existing overview system requires Overview Cache objects to speed delivery of NNTP XOVER requests to clients. Overview Caches store overview data in a format that can be streamed to a client connection. Without well-tuned Overview Cache, overview requests must lookup the data, then format it, then deliver it to the client.

Tornado implements a totally redesigned Overview system specific to Tornado alone. The new system bears the following features: