I am working on a project. I will need to live track 10.000 vehicle. Every second each vehicle send his localization. So Every second I will need to save 10.000 record. At the end of the day will have save 10.000 * 86.400(second/Day) = 864.000.000‬ record.

With the Data for each vehicle, we will like to be able to export between two date all the localization Data in GPX format, so it can be upload on GoogleMap or other similar tools.

I have supposed that the time between two Gps fix was one second. That seem to not be true.

So I have tried to find the average time between two GPS fix (average Time for GPS listening to send two localization). I wasn't able to find clear answer. I was only able to find information on the Time-to-First-Fix, knowing the answer will help me.

Because the quantity of data may be big I wonder if I should use a NoSQl DataBase? Or do you think relationship data base is enougth?

  • 4
    I would suggest not storing the data if it’s identical to the previous entry. This could save a lot of space, at the expense of making queries slightly more difficult. With only 10000 vehicles you could keep the current positions in memory for comparison — or better yet: move that check to the reporting vehicle. Each vehicle only reports a new position when it has new data to report. NB taking a GPS fix takes a long time so storing a location every second is meaningless — or something is already interpolating the raw GPS data. Nov 8 at 16:32
  • Good point that @Colin'tHart makes; if a vehicle goes 60 Miles per Hour, that is just one mile per minute; so in one second, only 87 feet per second. So maybe whoever is gathering requirements needs to re-think the goal. Are those vehicles spread across several jurisdictions in several countries? If so, a potential help would be partitioning the data by jurisdiction, or time, ... Nov 8 at 18:22
  • What use will be made of this data? That's more likely to determine the product used. Relational / nosql refers to the external interfaces and system guarantees, not the underlying storage capability. Nov 8 at 22:10
  • Number of data points is not a valid criterion for determining when to use NoSQL vs a relational database system, both are essentially equal in that regard, independent of how much data there is. More so, how the data will be consumed and the type of querying you'll need to do against it, in addition to if the data is being received in a structured and / or normalized form, and if it's already naturally related are some of the factors one uses to determine which type of database system to use.
    – J.D.
    Nov 9 at 0:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.