CODEWITHSUNDEEP
architecturesearch-engine
Vitali Pom
Vitali Pom

Reputation: 602

How does google return so many search results?

When googling for some term we're told that we see the k's 10 results out of thousands. I.e 1-10th results out of 10000. We can skip right to 50-60th result and google will return them in less than a second. If google would cache all of the 1-40, I guess it would waste RAM. If it would return all of the results at once, it would waste traffic. How does google return so many search results while it sorts them?

Upvotes: 2

Views: 1790

Answers (1)

rahulserver
rahulserver

Reputation: 11205

Here is something I tried: For the google search:

How google search engine works

Here is the url sent to the server:

https://www.google.co.in/?gws_rd=cr#gs_rn=20&gs_ri=psy-ab&tok=v1e-J_ynDKaPw_oDlFH-HA&cp=30&gs_id=8&xhr=t&q=how+google+search+engine+works&es_nrs=true&pf=p&output=search&sclient=psy-ab&oq=how+google+search+engine+works&gs_l=&pbx=1&bav=on.2,or.r_cp.r_qf.&bvm=bv.49478099,d.bmk&fp=56e0e38cda57afa7&biw=1280&bih=899

And if I select the page number 5 link on the bottom of the page, here is the url which is being sent:

https://www.google.co.in/?gws_rd=cr#q=how+google+search+engine+works&ei=qm3qUeXSA4LqrAfXtYDoCg&sqi=2&start=40&sa=N&bav=on.2,or.r_cp.r_qf.&bvm=bv.49478099%2Cd.bmk%2Cpv.xjs.s.en_US.c75bKy5EQ0A.O&fp=56e0e38cda57afa7&biw=1280&bih=899

Note the additional start=40&sa=N parameter here.

So what I feel after going thru google search enging working step 2 of 3 that google has its own index for over 30 TRILLION web pages and for each query they rank the result using over 200 factors. So I feel that the start and sa parameters form one of those factors while fetching the result.

Hence my opinion is that google is not caching the results, but carries out a new search for each queries using different url params.

EDIT:

To further support my observation, here is a list of time taken for navigating to the different pages of the above search query:

  1. For first page: 157,000,000 results (0.22 seconds)
  2. For page 5: Page 5 of about 157,000,000 results (0.26 seconds)

  3. For page 10: Page 10 of about 157,000,000 results (0.30 seconds)

  4. For page 14: Page 14 of about 157,000,000 results (0.33 seconds)

  5. For page 18: Page 18 of about 157,000,000 results (0.38 seconds)
  6. For page 22: Page 22 of about 157,000,000 results (0.35 seconds)
  7. For page 26: Page 26 of about 157,000,000 results (0.43 seconds)
  8. For page 30: Page 30 of about 157,000,000 results (0.45 seconds)
  9. For page 34: Page 34 of about 157,000,000 results (0.40 seconds)
  10. For Page 38: Page 38 of about 371 results (0.50 seconds)

So the time is almost increasing for increasing page numbers of queries. If it were caching the results, above would not have been the trend.

Upvotes: 3

Related Questions