The filings with the United States Patent Trademark Office are all defined by the concept of "container" patents -- so called because they describe a computer within a computer concept for applications that use multiple servers.

Google expert Stephen E. Arnold, who has studied Google's patents since 2004, has done an analysis of the container patents. In a 277-page analysis entitled "The Calculating Predator," Arnold argues that the container patents discuss the way Google "is building and deploying a leapfrog solution in plain sight."

"The reality is that Google is going to be pulled into the enterprise, probably more rapidly than the company wishes. Once in the enterprise, the competitive backlash and industry reaction will be strong and mostly hostile." Arnold said in his paper

Noting that Google has already made some successful incursions into enterprise computing including its search-based features, e-mail, online calendars, and office productivity software, Arnold added that the firm's 8,000-plus Google Appliance population can be built upon to grow the container function in the future in enterprise installations.

"Google can use the appliance as part of the container functionality or implement the appliance as a virtual machine providing multi tenant services," Arnold said in an e-mail. "Software, not hardware, is where the enterprise division at Google is going."

The cited patents include USPTO number 2007 0136443 "Proxy Server Collection of Data for Module Incorporation into a Container Document" by Adam Sah, Dylan Parker, and Christopher H. Rohrs; number 2007 0136337 "Module Specification for a Module to Be Incorporated into a Container Document" by Adam Sah, Dylan Parker, and Christopher H. Rohrs; and number 2007 0136201 "Customized Container Document Modules Using Preferences" by Adam Sah, Dylan Parker, and Christopher H. Rohrs.

The so-called container inventions, Arnold said, constitute a mechanism whereby Google can place customized workspaces in front of employees. The Google container is actually a computer within a computer, according to Arnold's interpretation of the patents; the technology makes use of multiple servers that can deliver different applications.

The entire approach is built on XML. "The lingua franca of the system is XML," Arnold wrote. "What's clear is that Google has settled upon XML as the mechanism for passing data that has structure. Within an XML file, Google can place data, metadata, instructions, and nest or stack these instructions." XML eliminates the need of using vendor-specific proprietary scripting languages.

Google's recent patent applications increasingly emphasize the importance of XML for describing data, passing metadata, and issuing instructions for the various components of the Googleplex (Arnold's term for Google's massive network of server-based data centers).