Looking into The Google Legacy.

Navigation: Main page

Author: Hane, Paula J. isphane@infotoday.com

Is Google the next Icarus or the next Microsoft? Will the 7-year-old, highflying darling of Mountain View, Calif., fall down to earth when the hot rays of competitive reality melt its wax wings? Or, will it he the heir to the worldwide dominance wielded by Microsoft (which in turn displaced the mighty IBM)? Whatever you think its fate will he, Google is certainly not just a search company, as search expert Stephen E. Arnold clearly shows in his e-book The Google Legacy, which was published in fall 2005. Arnold contends that Google is greater than the sum of its parts. In fact, he says Google represents a disruptive technological force--one that has become a metaphor for the next generation of computing.

A sought-after consultant, popular lecturer, and established author on technology, Arnold has authored six hooks and more than 50 articles. He is the author of the popular Enterprise Search Report (published July 2005 by CMS Watch; http://www.cmswatch.com/Search/Report) and spent nearly a year and a half researching and writing The Google Legacy. I interviewed Arnold in late November 2005 to talk about the book and his thoughts on what has happened since its publication.

Arnold said the work for The Google Legacy began while he was preparing reports on various pieces of the Google empire for clients during 2004. He said: "I kept wondering how Google could deliver such computationally complex processes at such high speed. It puzzled me. So, I started poking around." Arnold read some IEEE papers by Google engineers; the footnotes led him to some ACM documents by Google professionals, and he started making lists of the company's engineers. He and a team of researchers then systematically tracked down Google's patents and read them.

Google didn't officially cooperate with Arnold in his research. In the hook, he writes: "[T]he information in this monograph comes from conversations with people who follow Google assiduously, Google's own white papers and patents, and authors of formal papers and informal Web logs who often provide tantalizing glimpses of the inner workings of Google." When people heard Arnold was researching Google, some forwarded technical documents and even videos to help him out.

Many Google employees (not to mention the employees at Microsoft and Yahoo!) were also interested in Arnold's findings--some even suspect they were eager first readers of the work. Arnold said that individuals he talked to at Google often knew only one piece of the overall Googleplex structure. Through his investigative research, Arnold managed to pull chunks of the company's diverse technology pieces together.

The Google Legacy is a serious book. If you want to read about the food and the lava lamps at the Googleplex or the history of when the company's founders, Larry Page and Sergey Brin, met, this is not the hook for you. Instead, you should read one of the several interesting and entertaining books about Google that were published in the past year, such as The Google Story by David A. Vise and Mark Malseed. "Mine is not breezy, anecdotal, and easy to read," said Arnold. "Mine is a conceptual analysis of [Google's] hardware and software engineering and the technology that it supports."

As the press release of The Google Legacy accurately states, the book "places Google under a microscope, dissects Google's technology, evaluates its potential, and determines that Google's future lies beyond search."

The details Arnold managed to assemble about Google's operations (its data centers, hardware components, server racks, network architecture, redundant file copies, etc.) are remarkable. And, he shows a thorough grasp of the key differences between Google and its competitors. Yahoo! has grown through acquisitions and has a mosaic of different servers and systems. Microsoft, primarily a software company, has a "legacy code boat anchor" around its engineers' legs. Perhaps the best part of The Google Legacy is that it's not overly geeky, dull reading. Despite what he might say, Arnold tells a good story, with useful analogies and clear writing.

As Arnold details, Google's computing platform (named the Googleplex by Arnold after the name given by the company to its Mountain View, Calif., headquarters) is a better (i.e., faster, cheaper, and simpler to operate) computer processor and operating system than systems now available from competitors. The sum of Google's innovations is the Googleplex, which Arnold sees as "the next-generation computing platform."

And the differences are dramatic. Arnold states in the book:

If one does a rough calculation of the cost advantage that Google enjoys, the difference between the costs of a branded Unix system from IBM or Hewlett Packard or even a Windows system from Dell is that Google enjoys at least [a] four to one advantage, maybe more. This means that for every dollar Google spends, a Yahoo or Microsoft would have to spend four, just for hardware. In terms of some system maintenance and administration chores, Google enjoys an even greater advantage because its costs are effectively zero due to its innovations in operating systems for massively parallel, distributed systems.

And, according to Arnold, the proof is in the service: "The company has suffered no significant downtime since 2000."

Here's how Arnold sums it up mathematically: "The recipe at Google is surprisingly simple: Smart People + Googleplex + Data + Cool Algorithms = Competitive Advantage."

Pointing out the synergies of Google's two founders, Arnold said that Page is a hardware expert whose father is a world leader in distributed parallel computing. Brin's expertise is in algorithms and his parents happen to he leaders in that field. So, these two partners, with "genes and genius, glued together the best bits and pieces from research computing and solved a couple of fundamental problems that plagued Internet search companies like AltaVista and Inktomi. Then, as they refined their engineering between 1998 and 2004, they ended up with much more than just search. Google became an applications platform."

Arnold has a way of summarizing the complex in a concise, understandable way:

Yahoo has smart people, essentially the same algorithms, but less well-integrated user data. What is missing is the Googleplex and its data consistency. A similar situation exists at Microsoft: smart people, disparate data, and an older-style computing infrastructure. Other companies have hardware, software, computers, and brains stuffed with mathematical [know-how]. … Compared to these [search] competitors, Google seems more focused and adept at squeezing the last few grams of performance from its infrastructure.

Arnold described the performance this way in our interview: "Goggle has a pretty good NASCAR, while Microsoft has a nice Chevy Impala--it's comfy to ride in but it just doesn't go as fast."

So, said Arnold, Google is this era's "transformational computing platform" and could be about to unseat Microsoft from its throne. While he made it clear that it's too soon to judge if this will actually occur, he cautioned that we should not view Google as just a search company that sells ads. "What is now being recognized is that search is just one application running on the Google computer. The real secret sauce at Google is its robust technology that can morph at a moment's notice and catch its competitors off guard."

So, Google does not have a search system. What it has is a supercomputer that delivers applications--it's really the network computer for an increasing number of users each day. "Google did not invent distributed computing. Google simply figured out how to combine hardware and software engineering to create a low-cost, high-performance virtual applications delivery platform," Arnold explained.

Arnold sees Gmail and Google Maps as true, compelling virtual applications that provide Google with the levers to break Microsoft's grip on the desktop. Arnold said, "Gmail is the first Google application that hints that Google technology can deliver a network-based file storage system, maybe a word processing program."

While Google has been releasing what might seem to be fairly diverse, separate applications, the company actually is starting to assemble the pieces into a bigger picture. Not only can its products be put together and grow into something that resembles Microsoft Office, but they could expand into many other applications, for example, a music booking service for local bands.

And, with the expected growth in mobile computing, Arnold said these network-accessible applications will become even more widely used. Other avenues that Google could conceivably pursue, besides word processing applications, might include an alternative to Yellow Pages for small business advertising, a presentation application that would challenge PowerPoint, and a contacts function that could match features available within Outlook. Google could even choose to add reports and features to provide a customer relationship management application similar to Salesforce.com's service.

Furthermore, Arnold noted that Google has been acquiring fiber to connect its distributed data centers, which each have as many as 10,000 servers. "We could be moving to GTT--Google Telephone and Telegraph," Arnold said.

Google is in the process of transforming enterprise search just as it did the Web search market, according to Arnold. This is all about generating additional revenue. Google's recent quarterly financial filing confirmed that 99 percent of its revenue comes from online advertising. But, as Arnold noted, even 1 percent of $6 billion is a big number. Google already has 2,000 enterprise customers and is gaining more all the time.

Google has made it clear it's serious about the enterprise market. Following the recent announcement of the Autonomy acquisition of competitor Verity, Google announced a "search replacement program" whereby any business that replaced an existing Autonomy or Verity search solution with a Google Search Appliance would receive a free Google Mini (worth about $3,000).

In his book, Arnold writes: "With a hardware presence in the server room and a desktop software presence in the enterprise, Google has the foundation in place on which to deploy additional enterprise services. On the surface, Google's strategy appears to be winning control of the enterprise search market. In reality, Google may be taking the first tactical actions in a wider campaign to create a platform for deploying other, more sophisticated enterprise applications on the search foundation.

"Google Earth is another clear example of Google's larger strategy for expanding its presence in the enterprise," he continues. "Just as Google Search buzz masks the importance of Google's enterprise search initiative, the Google Earth as a type of purposeless online game obscures the enterprise impact of the Google GIS initiative. Search and Earth are beachheads in a larger battle for enterprise revenues."

Arnold called Google's introduction of two fledgling services for its users ( Google Print and Google Scholar) "possibly ill-conceived and poorly integrated." Since he finished the book, a lot has happened, particularly with Google Print, which has been renamed Google Book Search (another slap at Microsoft, which is working on its MSN Book Search project). Yahoo!, the Internet Archives, and many companies and libraries have partnered in the Open Content Alliance to develop the Open Library, pushing the notion of open standards and access.

In Arnold's view, since Google didn't implement things quite as elegantly as it wanted, the winner is the user. He feels that, despite the push to open standards, Google will probably continue on its path of creating its own collections. But, it's no more mainstream an effort than anything Google is currently testing, in his estimation. Google is capable of just shifting its energies if something fails to get traction. "It's such a minimal effort for Google to try new things that it can try this Darwinian approach."

Arnold obviously hit a hot button, coming out with his book in late September when Google seemed to be dominating the news (though when hasn't it, lately?). The technology press picked up on it immediately, with articles highlighting Arnold's work featured on CNET, ZDNet, eWeek, Information Week, and more. He's also been quoted by the Associated Press, USA Today, and many technology bloggers.

Veteran search and content management expert Martin White reviewed The Google Legacy for Free Pint. He had high words of praise for Arnold's achievement: "I thought I knew a fair bit about Google, but it is the tip of the iceberg compared to the knowledge that Steve Arnold has amassed and presented with great clarity in this very highly recommendable book. … This is not a journalistic approach to Google but the outcome of the author's lifetime involvement with search applications. The result is a level of technical detail and analysis which I cannot see ever being bettered." Putting things in perspective. White reminded readers of fairly recent Internet history: "Equally valuable is that Steve Arnold looks at some of the issues that might yet derail the Google train. After all, I can remember the days when no one could conceive of there being a competitor to AltaVista."

According to Arnold, "Google is faced with a number of legal, management, and revenue challenges. Any one of these can reverse Google's upward trajectories in revenue, users, and traffic. Combine two of these factors at a propitious moment, and Google can slip into a death spiral."

I asked Arnold for his best guesstimate on Google's chances. He admitted that the sheer newness of Google's architecture and its use of commodity code could be a stumbling block, as could the issue of click fraud, which erodes confidence. A lawsuit could also shut them down. Or, perhaps a Google-like competitor could emerge from China. He predicted: "There's an 85-percent likelihood that the company will continue unimpeded for at least 12 to 18 months."

If Google manages to survive all these potential pitfalls, it could become the next Microsoft. "If this happens," Arnold said, "users will benefit. If Google fails, users will also benefit because the legacy of Google is to make clear a better way to perform certain types of work, deliver information, and engineer a system." And, Arnold stressed, if Google somehow fails, its smart employees will just be hired by another company to work on the next Google. The point is that Google has shown the way to a new computing model--Arnold calls it a "phase change" (see table).

Some have suggested that Arnold presents a dark and ominous prediction of a monopolistic Googleplex--one that should be feared. But, his view doesn't seem to be that dark at all. He writes: "Google's legacy will persist even if Google goes out of business. Google's legacy is a light emitting diode illuminating a new computer, software, systems and information landscape with a cool, white light." In his conclusion, Arnold sums it up neatly: "Google's legacy is that it has shown the way to the next-generation computing and information platform."

The Google Legacy by Stephen E. Arnold is available in printable PDF as a Web site download (24.5 megabytes) only from Infonortics (http://www.infonortics.com). It is approximately 290 pages. The price for a one-time download is $180 (€145). Passwords and access codes will he communicated via e-mail. Arnold said that a PDF epilogue to the report will he available for free to all purchasers early this year.

Arnold Information Technology (AIT; http://www.arnoldit.com) has also released a CD-ROM containing the full text of more than 120 patents it believes have been assigned to Google or are held by key Google employees. This disc covers the period from the company's inception in 1998 through July 5, 2005. It is available for $50 plus shipping and handling. AIT issued the CD in response to requests from purchasers of The Google Legacy. An updated disc covering patents from July 6, 2005, to Dec. 31, 2005, will be available for free to purchasers of the original CD early this year.

Legend for Chart:

B - 1964-1984
C - 1984-2004
D - 2004-2024

    A

                B                  C                  D

Computing
Approach

         Glass House with    A Computer on    Network
         Raised Floor        Every Desktop    Ecosystem

Applications

         Batch               One PC, Local    Virtual
                             Applications     Applications

Work Mode

         Do not bend,        Tele or Flex     Anytime, Anyplace
         fold, staple,
         or mutilate

Technology

         Mainframe           Personal         Network
                             Computing        Computing

Users

         The Super           The Affluent     Almost
         Affluent                             Anyone

~~~~~~~~

By Paula J. Hane

Paula J. Hane is Information Today, Inc.'s news bureau chief and editor of NewsBreaks. Her e-mail address isphane@infotoday.com.