Where can I find information on how to implement the GPS monitoring service?

Question

Python

There was a task of writing own system of monitoring of motor transport. I must say right away - of all the found, both open-source and proprietary services, not one is suitable. Even the most flexible one will have to finish and rewrite most of the functionality. According to this, it was decided to write from scratch.
But in order not to walk through the field of rakes, I decided to appeal to the collective mind for materials on this topic. Interested in design nuances, some life hacks, etc.
Now I have already added a server python daemon for collecting data from trackers, parsing them and outputting them to the database. The demon goes through all the circles of hell in testing and finishing. There seem to be no problems here. But there are problems in the design of the future database, because according to preliminary estimates, the database will increase by ~ 1.5 million records per day. So far, there is an idea to store the parsed "raw" data for 8 hours, "archive" the rest hourly and add it to another table, where they will lie for another month. Everything that is older than a month goes to a long-term archive until better times.
- Why 8 hours? I think this is quite a logical time period for operational monitoring, without using samples from archive tables.
- Why a month and more in a long-term archive? For the construction of monthly reports by the dispatching service.
But still, something tells me that this is never the best option. Send please. in the right direction. I don't want to suffer from inventiveness with a waste of precious time.
PS.
- server OS - Ubuntu 12.04 Srv x64
- data collector - python-daemon for version 2.7
- base - MySQL 5.x (perhaps not the best solution for this task)
- frontend - PHP + JQuery + Apache
Maybe I missed something in the story so far wrote, well so I will specify.

Answer 1

MySQL in my opinion is not the best solution in the forehead. Probably there are more adequate databases that:
1. perform data archiving
2. access data by time
3. access data by location (is this necessary?)
In general, what requests do you plan to process in the future?
I won’t say for GPS, but your task is related to the task of storing raw archives from the stock exchange (look here ), or storing statistics (for example, the modified solution above - see here , or here is a large list of databases, you can find out what is there). There, though, sampling is possible only by time period, but probably most people need it.
After all, why not write it yourself? Let's try? ;)

Answer 2

What services did you consider?
We at cloud.doroga.tv have already decided a lot of things... and in terms of data collection, storage, analytics and web interfaces...
write if you are interested

Answer 3

Try to collect raw data in RRD - a fixed size ring database, ideal for collecting metrics. There is a binder for Python. After that, you can consolidate/archive the data, for example, once an hour/day, and store it in NoSQL storage.