Google Search Appliance software version 6.0
Posted June 2009
The Google Search Appliance enables you to provide universal search to your users. You can get the most from your Google Search Appliance by using some or all of its many features to fine-tune and enhance universal search. Become familiar with the Google Search Appliance's features by reading this document and apply those features that best suit your search solution.

Before you launch your search appliance implementation to your entire user base, test it. This section contains some guidelines for performing the following types of testing:
After you have developed a universal search solution, run tests to be sure that it will have acceptable performance under production loads. Some guidelines that you might follow for performance testing include:
Also, when measuring serving performance, consider both throughput and latency, because both have an impact on the experience of your users. Throughput and latency are closely related. The search appliance has a fixed number of threads for processing requests. If your load testing script uses the same number of threads to send queries to the search appliance, you can calculate the maximum throughput given the average latency or vice-versa.
For example, suppose there are five threads continually sending queries to a GB-1001 and you see a throughput of 1200 queries per minute. Average latency would then be 0.05 seconds per query (60 / 1200).
In many cases, it is more meaningful to know the median latency or the maximum latency seen by the fastest 90 per cent of all queries.
For some of the most popular queries, find out what users believe to the be most relevant result. What is the actual top result in the search appliance? Relevance ranking is best done by users, because administrators can have a different viewpoint on the most relevant result.
To understand why a document is considered relevant for a search query, you can look at the context of the search terms in the document. If the search terms are in a header or title, for example, then the document is likely to be more relevant. Take note that search terms in your query may be expanded by the search appliance to include related queries.
If you have used search appliance features in your universal search solution, ensure that your testing includes those features. For example, if you have implemented source results biasing for specific URLs, examine search results to determine the effects of the strength and influence settings that you have chosen.
Once the content has been indexed and you have used front ends a features to create the appropriate search experience, you might deploy universal search to a small group of users for testing. This way, you can find issues and fix them before the entire user base begins to use the system.
After you have tested your search appliance implementation, you are ready to go live with your universal search solution.
When you launch your search solution, you can raise user awareness and excitement about it by sending periodic email announcements during the weeks leading up to the launch date. In emails, you might promote your search search solution as the exciting new resource in the enterprise for finding all types of information. You might also include tips and other information that people will find helpful for using the new search solution.
To help stimulate interest, enthusiasm, and familiarity with the search solution, you might consider holding a contest among users for:
Once people start using the search solution, they will trust that they are getting the most relevant results for their searches. Based on user trust and satisfaction, search traffic will increase over time.
As search traffic increases you can monitor it and identify ways to improve the search experience and fine-tune your implementation. These improvements can help increase search traffic and the effectiveness of your search solution for your enterprise even more.
Let Google know what you think about this document by sending feedback to gsadoc-gtm-feedback@google.com.