Skip to main content

A Giant called Google...



    Wondering what you will be reading in this post today and what is "Google Giant"! Well, I was working on a simple project a couple of months back and that is the time when I was amazed and by the size of data that Google Maps servers might be having! As most of the readers (developers and enthusiast included) would know or might have noticed, Google maps are rendered by stitching small images together. You must have observed this on quite a number of occasions, when you see grey tiles when you have a slow internet connection in particular. It takes time for the images or tiles as we call them to be displayed.

    Let us try and estimate what could be the size of the Google Maps data! Before proceeding, let's make a few assumptions and considerations.
1. Let's consider the Google Street View maps only.
2. The zoom level that is considered for the calculations is 21 - the max zoom level allowed on Google Maps API.
3. All the numbers are just an estimate and not actual values.

    Google maps zoom level goes from 0 to 21, displaying 1 image at zoom level 0 and 4,398,046,511,104 (about 4 trillion i.e. 4x10^12) image or tiles at zoom level 21. 4,398,046,511,104 tiles at zoom level 21 represent the entire world surface which is about 510.072 million sq km (196.940 million sq mi). Lets say, about 70.9% of the world is water and 29.1% is land, which would mean about 3,118,214,976,373 tiles would all be blue in color. About 30% of the land area is under forest cover, which means about 383,949,460,419 tiles would be green in color. The remaining 895,882,074,312 will have multiple colors depending upon what part of the land is part of these tiles. If the numbers have gotten too large, let's tabulate them.

Sr. No. No. of tiles Color of tiles
1 3,118,214,976,373 Blue
2 383,949,460,419 Green
3 895,882,074,312 Multicolored

    Now that we have a clear picture of the number of tiles and the colors they would have, let's try and estimate the size of these tiles. Each tile or image is 256x256 pixels and each blue and green color tile/ image would then be around 4KB each. This means that the total size in KB of all the green and blue colored tiles/ images would be about 14,008,657,747,168KB; that's about 13,046.5TB. This is huge and still there are about 0.8 trillion tiles left whose size needs to be calculated.

    Let us go ahead and do that as well. Since these remaining tiles are multicolored, the size of each tile ranges in between 13 to 17KB. For our calculations, let's assume that the size of each of the multicolored tile is about 15KB. That would mean that the total size of these tiles is about 13,438,231,114,680KB; that's about 12,515TB.

    That means; if my math is not wrong; the size of all the Google Map tiles for Google Map Street View at zoom level 21 alone is about 25,500TB.
 
    This 25K TB space is required to save tiles only for one type of Google Map at the highest zoom. There are tiles at other zoom levels and also tiles of other map types - Satellite Map and the Terrain Map. The labels that appear in the Hybrid map type is also a separate layer on top of the Satellite map. That means about 25K TB more for the satellite map at zoom level 21 and about the same for hybrid map at the same zoom level. Plus about a 15K TB for the labels layer in hybrid map. If the approximation is to believed, Google has around 100 thousand TB space for the maximum zoom level of all the map types.

    Yes, it seems comical to attempt such a calculation where every step of the way requires an approximation. That's why in the end I end up with a such a wide chasm of error. And, of course, this rough analysis does not cover area distortion introduced by mapping the globe's points onto a two dimensional surface. But, there you have it - the story of "Google - A Giant".
Labels:

Comments

  1. BharathiApril 12, 2012 at 3:28 AM

    Nice stats.. :)

    ReplyDelete
    Replies
    1. Shreerang PatwardhanApril 12, 2012 at 4:20 AM

      I loved them too...I had those in mind, but when I actually put them in ink today, I too came to appreciate them all the more...:)

      Delete
    2. blogaloloApril 12, 2012 at 12:32 PM

      I can die ...happy to know that eternity exist! thank's Google

      Delete
    3. AnonymousApril 15, 2012 at 11:29 AM

      Excelent investment. Do you have considerer include at your job BI since an aplication GIS. Today exist the option BI include GIS but not GIS include BI.

      I wait that very soon you reach your Master

      Member GIS Linkedin

      Delete
  2. A. BelienApril 16, 2012 at 12:10 AM

    Exept... Google probably reuses the same blue and green tiles for all water and forests. So, this means the 13,046.5TB calculation you made is actually just 8KB. The total number of storage then drops to about half. =)

    ReplyDelete
    Replies
    1. Shreerang PatwardhanApril 16, 2012 at 12:37 AM

      What you are saying is quite possible Alex. Bu I still think there would be several copies of the same blue or green image that would be rendered. But yes. The size would definitely be diminished.

      Delete

Post a Comment

Please leave your comments here...

Recommended for You

Where does Google get it's live traffic data from?

Referring to a post that I wrote earlier, Google’s - Live traffic Layer , ever wondered how Google collected this data? I was wondering the other day, how Google received live data to display it on their maps as a layer! I looked up the web and found something very interesting and am sharing the same with you all. As we all know, the traffic layer is available most accurately in several states in USA. Most major metro areas in the US have sensors embedded in their highways. These sensors track real time traffic data. Easy to miss at high speeds (hopefully anyway, traffic permitting), more commonly noticed may be the similar sensors that often exist at many busy intersections that help the traffic lights most efficiently let the most amount of people through. The information from these tracking sensors is reported back to the Department of Transportation (DOT). The DOT uses this data to update some of the digital signs that report traffic conditions in many metro areas. They als
17 comments
Read more

Geodesic Polyline

    Today we will have a look at a very interesting polyline example - "The geodesic polyline". Now the first question that will pop is "What is geodesic?". Mathematically, geodesic means the shortest line between two points on a mathematically defined surface, as a straight line on a plain or an arc of a great circle or sphere.     The next question after reading the above definition is clearly, "Why do we need geodesic polylines?" and that would be followed up with "What is this Great Circle?". We will discuss this first, before we move on to the actual example today. The example is very very similar to the normal polyline example, with just a small change.     Having said so, I will now try to explain why we need a geodesic polyline? The shortest distance between two locations on the earth is rarely a straight line as the earth is roughly spherical in nature. So any two points on the earth, even if they are very close lie on a curve and n
5 comments
Read more

Playing with the markers and info window bubbles...

    In the last few posts, we have seen some marker examples and some information window examples. Now, lets do something interesting combining these two things. Just writing that "This is an info window" in the information bubble is not very interesting! And I know this...Have gone through the same phase!     So, today we will do something interesting! We will display the latitude- longitude co-ordinates of the point that the user clicks on the map! Doing this is not at all complex! Copy paste the following code and you will see for yourself a map coming to life!     The output of the above code looks as seen in the result section above! If you have any queries regarding the above code please comment on the blog post or feel free to contact me at my mail ID .
3 comments
Read more

GIS Technology to identify all properties in BBMP Limits

          The Bruhat Bangalore Mahanagara Palike (BBMP) has for the last two years, been in the process of conducting a massive exercise to map the 13.87 lakh properties in the 198 wards in the city. Geographical Information System (GIS) has proven to be an effective tool for analysing and displaying thematic maps of all the roads for proper evaluation and correction of zones.            As of now, 11 lakh properties have come under the tax net. The aim of this programme, which is perhaps the first such exercise being conducted in the country on such a large scale, covering 800 square km, is to bring all properties under the tax net and ensure that the BBMP has accurate information pertaining to the properties’ dimension, built-up area, land use and classification. The process of validation of GIS maps will be completed in January 2011.           The process uses satellite digital maps of the BBMP area to generate a vector map from the information obtained. These high r
5 comments
Read more

The bitter divorce of PSD and HTML

    Today's article is an interesting post that I read. The original post in Portuguese and authored by Fabricio Teixeira  can be found at arquiteturadeinformacao  (Now don't ask me pronounce this =)).     Some are calling it the death of PSD  but I prefer calling it a "divorce". PSD and HTML are both healthy and living strong, just that they do not live together anymore. "PSD to HTML", which for years was the most accurate and sometimes the only right path to web design process, seems like has its days counted.     Firstly you draw a page in Photoshop; impeccable layout, representing exactly how the web pages would appear when opened in a browser. After a sign-off on this picture (PSD) from the client the front end developer transforms these pictures into HTML, CSS and Javascript. The assets are cut, one by one, exported from the PSD and integrated into the HTML. Plugins and new tools are created in the process and some companies even charge upto $1
5 comments
Read more