URL Classification - Cache sizes and Frequent Updates

Introduction

The Frequent Update mechanism updates the engine and content, such as classification rules, heuristics and databases.

The URL (or mail) signature database updates are sometimes very large, and so they are not applied (stored) to the database straight away, but they are however used immediately.

The frequent update files are added to the cache, and stored to a directory for a scheduled merge process that runs later.

The cache is always queried first when an URL request is invoked. You may change the size of this cache by changing some settings in the related init-files (Database Connection Options).

At regular intervals the update files will be added to the database in a merge process. This merge process is also triggered whenever update files are added to the cache, and as a consequence the cache size exceeds the given maximum cache size.

Setting too small a value for the maximum cache size will lead to more frequent database merges. Since the merge process generates a lot of I/O traffic and also consumes some CPU time, you should carefully set up the used cache size and merge interval.

Recommended values are related to your application and are hard to state. Try out the values you have chosen for a test period to see how the updates and the merge process influence your application's performance.