The Web 2.0 Revolution (5) Search 3.0
Search engine itself is a product of Web 1.0; it is so not Web 2.0.
[+] Search is a product of Web 1.0
In my last article, I talked about the business paradigm for Internet media and online advertising brought forth by Google. It is about "many a little makes a mickle," which is very Web 2.0 in every way. Now it looks like Google will become the new paradigm which is going to last happily for ever. Yet, is it really so?
Let's not forget that search engine is a product of Web 1.0 and has been there for over ten years. The logic of search engine is to have everyone connected to a certain website, where there is a central host that does all the computation and comes up with search results. This is so not Web 2.0.
In addition, although Google has managed to enhance the accuracy of keyword search to an extraordinary degree, very often users still get a chunk of irrelevant results. The next generation search technology thus becomes the target eagerly pursued by all major Internet giants.
Many people, however, think that the weak link of technology can be compensated by "social search," as Web 2.0 is characterized by "sharing and participating by the public." After all, computers are driven by programs or a fixed logic, which can never keep up with the flexibility of human brains. As such, it seems to make sense to improve search accuracy by human assistance.
A practical solution is to let an Internet user define the keywords (the so-called "tags") for her/his own content when posting articles or pictures. This will make it easier for other people to find the user's content when they search by the same keywords.
[+] It is the search cost that matters
This is what is called "social search." By having people define the keywords or tags for their own content, and search engine index and compute these tags, the degree of accuracy can be further enhanced. After all, these keywords are defined by human brains and may be more likely to fit in with other people's way of thinking.
This search method - we may call it Search 2.0 - may not appear very relevant to technology. It is also the result of the ever-declining storage and bandwidth cost, which makes it possible for everyone to produce content on the Internet.
Ten years ago, it was unlikely for such a search method to emerge and diffuse, let alone to become the mainstream.
Yet "social search" is not the ultimate solution. In fact, practically all search methods will have to face the problem of
cost in addition to accuracy. This problem can be roughly represented by the following equation:
[+] Every search you make online costs money.
"Every search you make online costs money." In the past few years, the explosive growth of searches has been driving search engine operators to purchase new equipment and to expand the space for settling such equipment from time to time, so that operators can provide accurate and fast search service.
From the above equation, we can see that if an operator fails to control its cost properly, the average cost per search
will increase. However, if reduction of cost will sacrifice the performance of accuracy and response time, the average search cost will drop but users may start to leave.
[+] Towards Search 3.0
Thanks to the decreasing storage and bandwidth cost on the Internet, equipment is not as costly as before. Hardware size has been shrinking, which has made it possible to accommodate more appliances in the same data center.
Therefore, the cost of data center may be economized and the financial pressure on operators may be somewhat relieved. However, in spite of the many efforts, it has been very challenging to drive the computation cost down.
Google is one of the search engine websites which has a clear understanding about this point, so it has attended to the cost of equipment very carefully. It rarely purchases host computers from external vendors - tens of thousands of its servers were designed and developed by it and assembled from low-cost PC components. That is how it keeps cost down.
Still, Google needs to deploy data centers in order to provide better quality of service. Its new Oregon data center, built along the bank of Columbia River, is as big as the area of two football courts, and has all sorts of hosts and servers.
There is nothing wrong about such way of thinking; it is only very Web 1.0. No matter how deep the pockets of search engine operators can be, it can only own a limited quantity of servers. Be it ten thousand or one million, it can only be fraction of the hundreds of thousands of PCs globally.
I would like to reiterate the spirit of Web 2.0: "many a little makes a mickle." When a search engine user clicks on "Search", which mode do you think will be the faster to get what s/he wants? Through the Oregon Data Center in the U.S., or through a huge number of PCs all over the world?
[+] P2P search: the next generation search method
In fact, Google has developed many technical innovations including parallel processing, which is to divide a single query into several ones to be sent to several servers. This technology may have been applied to Google's search engine. Every search you make may involve several servers.
The question is, when will the above mentioned "several servers" become "all PCs connected to the Internet globally." Instead of Google's host, every search of yours involves thousands of home PCs, like the one you are using now, at different households the world over.
If one day the impact of technological innovations and the ever declining average cost of home broadband access and storage worldwide have got to a certain critical point, the existing search method on the Web may be overturned! Due to the technological bottleneck and security concern behind such expectations though, it is not easy to have significant breakthroughs in the short term.
In addition, there is also an economic problem: why do I need to offer my PC and home broadband connection to process your search? Maybe "one day I may need your help with my search" is a good reason, which sounds even more advanced than Web 2.0. We may as well call it P2P Search.
With the continual reduction of the cost of online storage and broadband access, will there be one day when it is cheaper to throw a query to PCs worldwide than to proprietary servers? This is an issue that search engine operators on the globe need to think over seriously.
- Read More
Prev : The Web 2.0 Revolution (4) the Google Paradigm
Next : The Web 2.0 Revolution (6) Struggle of the Press Industry
- Today in History
The Web 2.0 Revolution (5) Search 3.0 - 2006/09/24
Crime and Punishment of P2P (1) Liberalization of Power - 2005/09/25
Three Musts of Digital Content Biz (3) Redefining Ownership - 2004/09/26
Corporate Website a Handful (1) Accountability Where? - 2003/09/28
Recent Comments