Managing the Google Search Appliance | Google Power Tools Bible

The Google OneBox for Enterprise, shown in Figure 38.1, gives you access to data in 220 file formats and 109 languages, no matter where it’s located throughout your enterprise. This includes data residing in databases, on Web servers, and in business information applications such as Customer Relationship Management (CRM) and Enterprise Resource Planning (ERP) systems. That’s why it’s called OneBox. Using “one box,” you can type search queries for a wide variety of business information.

image from book
Figure 38.1: Install the Google Search Appliance for access to large enterprise data sources.

To understand the advantage of the Google approach to information retrieval, first think back to how information in a business was previously retrieved. Databases were queried using complex SQL (Structured Query Language) statements, or report writer software was employed to create endless ad hoc reports, sometimes to retrieve a single bit of information. There were no effective means to search stored documents unless they were stored in a managed data warehouse program that normally required regular system administration. The Google advantage is the ability to index documents in a wide variety of information storage mechanisms and the means for retrieving data in a unified manner, with zero end-user training. It takes no more skill to search through all your business data than it does to use the Google Web search box found at Google.com.

These are some of the other features Google is famous for that assist you in retrieving enterprise information:

Spelling suggestions: When you type a misspelled word, Google suggests the correct spelling. Because Google has indexed your business information, this works even with your business-specific terms.
Relevant results: The true power of Google searching is based on Google’s PageRank technology. Your search might result in thousands of possible business documents. Google helps you find the right information by suggesting the most relevant results first. Enterprise results also can be grouped based on where the information resides.

The Enterprise version provides even more services such as the ability to view a document without having the original application installed on your computer.

How it works

In the same way that Google Web Crawlers index billions of Web pages across the Internet, crawlers find their way through your business information, creating a master index of all your documents. Because many business documents contain sensitive data, the Google OneBox has an effective security management system that allows access to data only through secure permissions. Follow these steps to make a typical query:

The user (person who wants information) types a query into a search box provided on an internal company Web page. This is similar to the familiar Google Search box at Google.com. Instead, this box is customized with your company information.
The user’s query is analyzed to determine if a particular OneBox module has the relevant information.
A secure HTTPS (encrypted) GET request is sent to the relevant OneBox provider (information source). Security information is sent along with the request to determine if the user has permission to access the requested document. The communications are encrypted so that sensitive company data is not transferred openly across the network.
Results are sent back from the provider to the OneBox in XML format.
The XML is formatted into HTML by the XSL template located on the OneBox.
The results are displayed to the user in familiar Google results format.

Results are displayed using relevant snippets of the document that contain your keyword. Also, keywords are highlighted within the page to make finding the relevant information simpler.

Integrating Enterprise applications

One of the features of the OneBox search appliance over the lower-end Google Mini is its ability to integrate with your enterprise applications. The Google Search Appliance uses the REST-based API with a fully documented SDK for integration with enterprise applications. You create small programs called OneBox modules to access data. A OneBox module defines a type of search, the keyword used to invoke the search, how the information is obtained, and how it is returned to the user making the query. It also defines any security requirements for access to the data. You can write as many OneBox modules as you like. For example, LDAP, a form of hierarchical storage, is used to store a great deal of corporate information, particularly employee contact information. A OneBox module is written to access employee contact information in human resource systems. Remember, access to information in the Search Appliance can be restricted based on user permissions.

See many examples of enterprise applications in the OneBox Gallery at http://code.google.com/enterprise/oneboxgallery.html.

You also can access the full OneBox for Enterprise Developer’s Guide at http://code.google.com/enterprise/documentation/oneboxguide.html.

Managing the device

Management of the Google OneBox for Enterprise can be handled from any place on the planet where there is Internet access and can be managed in 16 languages. The administrator also can easily monitor the health (running status) of the equipment due to its SNMP (Simple Network Management Protocol) compliance. If something does go wrong, you can activate remote diagnostics so Google support can help you analyze the problem. While administration of the OneBox is simple, it also is powerful. The administrator can do simple things such as define synonyms for corporate terms so they appear as alternate suggestions when the user types a keyword. The administrator also can manage user access to information, see usage statistics, and monitor the health of the OneBox.

Managing accounts

You can manage user access, security, and permissions to the OneBox using a Web interface, and you can restrict access to documents based on custom security restrictions that you set.

Creating OneBox modules

Creating a OneBox module involves creating a trigger, selecting the provider, and formatting the output of the results. The OneBox can process queries all the time, when specific keywords are encountered, or when the query matches a regular expression.

There are two types of OneBox providers: internal and external. An internal provider returns information within the OneBox collection of information. An external provider is designed to retrieve real-time data from external applications such as databases, CRM systems, and other business applications.

Performing system analyses

You can view reports of user queries to help optimize the performance or your information system. This helps you quickly determine where the data is coming from, problems with access to information, errors that are encountered, and problematic servers. You can track system usage by the hour or by the day or track the use of certain custom features.