Google preprocesses its massive data set to be able to quickly filter down results according to your search string.
One simple way of doing this is to make a huge database of words, with links to every web page in which the word appears. Then when someone searches for a set of words, you can find the links that are common to all those words, and then from that much smaller list quickly return a ranked list of results.
This isn’t exactly what Google does, but it is analogous. The important part of the analogy is that this database, while smaller than the entire contents of the web, is still massive. Windows File Explorer could do this, but it would take time to create this and keep it updated, and it would take space to store it.
File Explorer [does do this](https://support.microsoft.com/en-us/windows/search-indexing-in-windows-10-faq-da061c83-af6b-095c-0f7a-4dfecda4d15a), but it balances search speed with other concerns. So depending on what you’re searching for and where in your filesystem you’re looking, you will not be taking advantage of this search index, and instead have to look through every file.
If I give you a few boxes of books and say “I think I put Dune in this box, can you check?” then it takes time to check each book and see if it’s the book I want.
If you’re a librarian, though, and I ask “Do you have Dune in the library?” the librarian can tell you quickly, because they’ve already gone and catalogued every book they have.
Google search is like that – the part most people see is the search page, but there’s a hidden part to their technology: their web indexing “spider” downloads, on a regular basis, a huge percentage of the web, and stores information about what the page is about, what keywords appear on it, and so forth, in a huge index. Then, when you search, the query is farmed off to a bunch of servers which each search part of that index and return pages that it deems are relevant.
How fast or slow something loads depends on two things, the size of the file and the transfer speed. When searching for something on Google you don’t load up all the websites you are shown, you’re simply going through a catalog of websites or images. Search results are sectioned in pages so you don’t even load up that many to begin with, and only load up more in chunks as you search more. Google is a search engine so obviously their core operation is focused on making searches streamlined. Their data centers and servers are immense, using very expensive and very fast equipment, and for searches the process is refined to require the least amount of data, so when you’re making an initial google search google servers go through their lists internally, which happens near instantaneously, and then produces a list on your screen that’s no more than a few megabytes and that has no trouble coming up quickly on most internet connections.
When you’re searching through your computer though the computer has to load everything itself, which on older system or cheaper hard drives may take time. Usually it’s not so much the specific performance of a hard drive causing slow downs but rather a weaker system that is strained through many concurrent processes. Most people’s computers are filled with junk programs and files that stack on top of each other and this causes the computer to run slow, because when you’ve maxed out RAM or CPU usage and you’re asking of your computer to search for a bunch of files it’s simply bottlenecking at other points and it takes time to sort that out and give you your results. Well running computers with updated operating systems and clean of junk with their systems not strained by other processes are not that slow to bring up results.
It’ not searching the web when you type something. It’s searching an index file, which is basically one big document.
The web searches occur when it’s otherwise not busy as it updates the index file.
Macs do something similar with their Spotlight search for computers. The computer will index the contents of the drives when not busy and when you search for something, it’s quickly looked up in the index.
Because google has 100000000000x the computing power your desktop or laptop has. It has entire servers towers (a server is basically a computer on steroids) dedicated to to indexing the web and creating keywords and related data. It does this process constantly every second of every minute of every day. Thousands and thousands of servers doing nothing but indexing and creating keywords. Your computer on the other hand has thousands of other things it also has to do just to be on. If you had a “regular” computer, almost half of all your processing power goes to simply running windows. Servers don’t use windows because they usually have one job, index. They have other servers to handle the other jobs, so imagine have a laptop for every process your computer engages in just to be a functional computer. One slow computer for every single thing. Security, file management, running programs…. Your house would be filled with computers doing one job.
As a web developer, just want to add my two cents. Not sure if Google still sells them, but in the mid 2000’s, we bought a Google Mini for our site. Every night, it would spider the entire site, in the morning it would send out a report on any broken links, missing images, etc, this was helpful for SEO, but it also indexed the whole site, when someone used the search box at the top of the page, it was using the google mini interface and found things very accurately, and very fast, now just scale that up to the web.
Secondly, about ten years ago I was developing an internal site for a large corporation, they wanted a, drop down display, results as you type, kind of thing. After every third letter of typing, it would “hit the server” to see if anything matched in cache, if so, send down the recordset and display in the dropdown, if not, hit the database with a %like% query, save the query in cache, send it down to the client browser, the cached recordsets were set to expire in an hour after the last use. It was fast for the clients but very taxing on the servers, it also got better as the day progressed, as the cache would be empty in the morning,
Having Windows spider your hard drive every night would not be too bad, but if you left your computer off during the night, would you want Windows to do it during the day while you are working? How much hard drive or RAM space are you willing to give up just to have a faster search that you don’t use that often?
Latest Answers