Google Responds to Jarg Allegation
January 15, 2008
Intranet Journal reported on January 14, 2008, that Google denies the Jarg allegation of patent infringement. I’m not an attorney, and claims about online processes are complex. You can read US5694593, “Distributed Computer Database System and Method” at the USPTO or Google Patents Service.
As I understand the issue, the Jarg patent covers technology that Jarg believes is used in Google’s “plumbing.” In The Google Legacy and in Google Version 2.0, I dig into some of the inner workings that allow Google to deliver the services that comprise what I call the Googleplex. Note: I borrowed this term from Google’s own jargon for its office complex in Mountain View, California.
If the Jarg allegation has merit, Google may be required to make adjustments, pay Jarg, or some other action. I have read the Jarg patent, and I do see some parallels. In my reading of more than 250 Google patent applications and patents, the key theme is not the mechanics of operations. Most of Google’s inventions make use of technology long taught in college courses in computer science, software engineering, and mathematics.
What sets Google’s inventions apart are the engineering innovations that allow the company to operate at what I call “Google scale.” There are Google presentations, technical papers, and public comments that underscore the meaning of scale at Google. According to Googlers Jeff Dean and Sanjay Ghemawat, Google crunches upwards of 20 petabytes a day via 100,000 MapReduce jobs. A petabyte is a 1,000 terabytes. What’s more interesting is that Google spawns hundreds of sub-processes for each query it receives. The millisecond response time is possible because Google has done a very good job of taking what has been available as standard procedures, blended in some ideas from the scientists doing research at universities, and advanced mathematics to make its application platform work.
Remember that search was an orphan when Google emerged from the Backrub test. Excite, Lycos, Microsoft, and Yahoo saw search as a no-brainer, a dead end of sorts when compared to the notion of a portal. University research chugged along with technology transfer programs at major institutions encouraging experimentation, commercialization, and patent applications.
What makes the Jarg allegation interesting is that most universities and their researchers tap U.S. government funds. Somewhere in the labs at Syracuse University, Stanford University, the University of California at Los Angeles, or the University of Illinois there’s government-funded activity underway. In my experience, when government money sprays over a technology, there is a possibility that the research must comply with government guidelines for any invention that evolves from these dollops of money.
When I read the original Google PageRank patent application US6285999, Method of Node Ranking in a Linked Database (September 4, 2001) I was surprised at one fact in this remarkable blend of engineering and plain old voting. That fact was that the assignee of the invention was not Mr. Page. The assignee was The Board of Trustees of the Leland Stanford Junior University. The provisional patent application was filed on January 10, 1997, and I — almost eight years later — just realized that the work was performed under a U.S. government grant.
I will be interested in the trajectory of the Jarg allegation. I wonder if any of the work disclosed in the Jarg patent has an interesting family tree. I am also curious about the various data management practices, generally well-known in the parallel computing niche, have been widely disseminated by professors teaching their students basic information and illuminating those lectures with real-life examples from the research work conducted in labs in colleges and universities in the U.S.
Litigation in my experience as an expert witness is a tedious, intellectually-demanding process. Engineering does not map point for point to the law. When the U.S. government explicitly encourages recipients of its funds to make an effort to commercialize their inventions, the technology transfer business got a jolt of adrenaline. Patent applications and patents on novel approaches from government-funded research contribute to the flood of patent work choking the desks of USPTO professionals. Figuring out what’s going on in complex inventions and then determining which function is sufficiently novel to withstand the scrutiny of cadres of lawyers and their experts is expensive, time-consuming, and often maddeningly uncertain.
Not surprisingly, most litigation is settled out of court. Sometimes one party runs out of cash or the willingness to pay the cost of operating the litigation process. Think millions of dollars. Measure the effort in person years.
As the Intranet Journal story says: “Google has responded to the patent-infringement lawsuit filed against it by semantic search vendor Jarg and Northeastern University, denying the parties’ claims of patent infringement. Google has also filed a counterclaim, asking the court to dismiss the patent in question as invalid.”
Will this be the computer scientists’ version of the OJ Simpson trial? Stay tuned.
Stephen E. Arnold, January 15, 2008, Noon eastern
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
