Park GT is a parking lot search service that utilizes crowd sourced geolocation information from its users to estimate the availability of Georgia Tech parking lots. It is designed as a web application with automaticparking detection to minimize user input required.
Finding available parking can be time consuming and costly. Park GT aims to give parking suggestions by using crowd sourced tracking of other users to determine the availability of Georgia Tech parking lots.
Open the project in IntelliJ. Configure the datasource and redis parameters according to your setup. If using our database, replace the password with the current one (Due to KMS rotates password every week, please consult us for a new password if you want to test our application). Build the application using the App configuration. Run the app to establish connection to the database. Run web application by installing the dependencies and start react web application using set HTTPS=true&&npm start (For specific instruction for launching react web application, please refer to this README. Open the application at localhost on the specified port Alternatively, connect via a mobile device by connecting to the device you are running the application on.
The Swagger service can be accessed at https://localhost:8443/swagger-ui.html# while running the application.
Park GT uses Redis to keep database information in memory to speed up response times. The parking lot API forwards GET requests to Redis which returns the appropriate values if there is a match for the requested key. If there is a cache miss in Redis, it fetches and caches the appropriate data from the database for future use. Additional nodes are added to Redis as necessary to store more data.
Using HTTPS, which is a combination of HTTP and TLS/SSL, the application's communications with the backend are secured to ensure confidentiality and integrity.
AWS KMS provides key management for Redis and the SQL server, allowing keys for them to be regularly rotated. We use KMS to rotate keys daily and keep Redis and our SQL server secure. CloudHSM further secures the system by physically isolating cryptographic keys on AWS from other users and meets the FIPS 140-2 level 3 standard.
The requirement from HTTPS protocol force every connection established from our application to be HTTPS connection as well. Therefore, all our database API query need to be through HTTPS connection. For example, our application would be able to query the database using the url https://localhost:8082/parking-lot but not using the url http://localhost:8082/parking-lot. However, HTTPS requires a CA certificate which we would not be able to get since it cost money and we don't own the domain of localhost. One solution that we tried is using a self-signed certificate. While it is easy to installed the self-signed certificate on the computer, it is really difficult to install it on the phone and for phone's browser to recognize this self-signed certificate. We have tried many methods but the phone keeps returning invalid certificate error. Therefore, one approach that we come up with to work around this issue is to using a remote storage API and for the computer to access and continously fetch from the database. The computer will then cache the fetched data in the remote storage, which the phone will directly access the cached data instead of through HTTPS connection. This way, we achieved database access from the phone while bypassing the limitation of HTTPS. This section of the code can be found in here.
Users hopping on Stinger buses after parking may cause the status to incorrectly update to driving despite already parking their car.
-Include additional information about situations when certain parking lots may be available for free
-Provide confidence levels for availability estimations based on estimated completeness of tracking data
-Reliably recognize users to continue using the same user ID in the database and be able to store their permit information
Waitz utilizes IoT sensors to determine how occupied an area is, allowing it to determine how busy study spots may be in locations it is deployed.