Programming for an embedded system can be sometimes hard to kick start due to the fact that it can be hard to find a compiler to start coding. When one installs a linux distribution it usually already comes with a toolchain available for the user to start programming. The toolchain is nothing than a set of softwares that enables the programmer to compile and debug the code. It is usually comprised of compilers, libraries, linkers and debuggers.
Recently there has been a big effort in trying to bring the Internet of the Things to the main stream. As an answer for this effort, Espressif decided to create a protocol for itâs famous line of low cost ESP micro-controllers with the goal of connecting sensors with the smallest footprint possible. The protocol is called ESP Now and it uses wireless IEEE 802.11 connection as the physical/datalink layer media.
Finding solutions for a NP problem usually involve a lot of computational time, and choosing the best time to stop the calculations and assume that a solution is good enough depends not only on the problem, but also on the application. In this report two different algorithms will be analysed in order to try to reach a conclusion about the impact of the computational time in the final result. The problem in question will be the Travelling Salesman.
For this project the requirements where that the station would be located indoors or outdoors fairly protected. So it should not be
waterproof like IP67(8) for example.
Some interesting measurements to have in a weather station are:
- Atmospheric Pressure
Docker is a tool used for running containers in the main OSes in the market. A container would be a slimed down version of a virtual machine. It came as an obvious next step of the movement of vitualizing everything after we reached a level of unused computing power on most machines. Most of the servers are over estimated during the desing phase of a project so that it can handle overload during the hard times. As a consequence, during most of the time those servers are idle and with very low memory usage. The intial response to that was to install hypervisors in the servers and load than with virtual servers that could be spawned and moved around easilly.