RELEASES
Latest Release
Version 1.5:
- Segregated the different functionalities of GioTTO between the CentralService and the DataService and porting these from the DataService to the CentralService and implement it
- New Acls, new permission level r/w/p for sensors
- Modularized code using Flask Pluggable views
- Moved All the new Rest APIs to the Central Service
- Moved all response constant strings to one place on DS
- Changed the way update objects are formed
- Modified Ui
- Time Series Graph
- Updated ML demo
- New Ui for the giotto demo
A Guide to Giotto: Getting Started and Details
GioTTO Description
Before you start working with GioTTO, take some time to read through our documentation and our presentation. GioTTO has four main elements: a CentralService, a DataService, Connectors, and a Machine Learning layer.
Each installtion of GioTTO for an IoT installation (crossing multiple buildings) has a single CentralService. The CentralService hosts the IoT metadata information, in addition to user information, access control lists, and structural information (e.g. number of floors in a building). The CentralService provides mapping between sensors and the DataServices where the data is stored.
DataServices store sensor time series data points from the underlying sensor networks. The DataService manages the sensors and those sensors' time series data points. A DataService may belong to any single administrative group that requires sole control over who can access the underlying sensor data points. The number of DataServices used is up to the discretion of our users; one DataService for a building, a floor, or individual subsystems are all possible configurations.
Connectors are the 'drivers' of GioTTO, allowing for communication between various IoT devices and GioTTO. GioTTO comes with a number of example connectors already written, and support for writting further connectors as needed.
The Machine Learning layer allows for supervised machine learning. A GIoTTO administrator or end-user can provide a label to apply to all of the sensor time series available in a particular location, for a particular segment in time. The machine learning layer communicates with the CentralService and appropriate DataService to train a classifier using these labels, and then can classify new input sensor data. This is all done without any manual feature extraction, training, or model tuning, allowing a novice end-user to create example-driven machine learning classifiers.
Before you start working with GioTTO, take some time to read through our documentation and our presentation. GioTTO has four main elements: a CentralService, a DataService, Connectors, and a Machine Learning layer.
Each installtion of GioTTO for an IoT installation (crossing multiple buildings) has a single CentralService. The CentralService hosts the IoT metadata information, in addition to user information, access control lists, and structural information (e.g. number of floors in a building). The CentralService provides mapping between sensors and the DataServices where the data is stored.
DataServices store sensor time series data points from the underlying sensor networks. The DataService manages the sensors and those sensors' time series data points. A DataService may belong to any single administrative group that requires sole control over who can access the underlying sensor data points. The number of DataServices used is up to the discretion of our users; one DataService for a building, a floor, or individual subsystems are all possible configurations.
Connectors are the 'drivers' of GioTTO, allowing for communication between various IoT devices and GioTTO. GioTTO comes with a number of example connectors already written, and support for writting further connectors as needed.
The Machine Learning layer allows for supervised machine learning. A GIoTTO administrator or end-user can provide a label to apply to all of the sensor time series available in a particular location, for a particular segment in time. The machine learning layer communicates with the CentralService and appropriate DataService to train a classifier using these labels, and then can classify new input sensor data. This is all done without any manual feature extraction, training, or model tuning, allowing a novice end-user to create example-driven machine learning classifiers.
Getting Started
Now that you've downloaded GioTTO and read about it, you're ready to go! Be sure to read through our Getting Started guide for GioTTO to make sure your GioTTO instance is set up in the best way for you!![]()
Now that you've downloaded GioTTO and read about it, you're ready to go! Be sure to read through our Getting Started guide for GioTTO to make sure your GioTTO instance is set up in the best way for you!
