Meteos is a Machine Learning as a Service in OpenStack.
Meteos allows users to analyze huge amount of data and predict a value by
data mining and machine learning algorithms.
Meteos create a workspace of Machine Learning via sahara spark plugin and
manage some resources and jobs regarding Machine Learning.
We are using OpenStack components like Nova, Sahara(Hadoop/Spark), Swift
etc. We are heavily relying on OpenStack Component because We understand
whatever we build, it should be openstack centric and give other project a
freedom to use the service easily for any other components.
Problem Statement -
1. Currently there are many ML libraries available and if user wants to
develop some applications they have to go to
installation/configuration/integration of component which they have to
2. To use one small API of one library, developer has to install complete
library. 3. 3. Processing primary data on cluster, required another
installation steps and managing analytics clusters is again a pain.
4.Storing and scaling data is an another major problem in data science
5. There are public cloud providers offering ML service but those are very
Meteos Project aims to provide Machine learning APIs to develop ML
algorithms and ML use cases with stack provided by Meteos setup.
This project gives one setup which includes Apache Spark
ML/Tensorflow/Scikit integration Hadoop/Spark and swift and it enables a
platform for developers to use ML APIs to develop ML use cases directly
without taking headache of installation and component integration.
*Planned features for Queens release - *
- Tensorflow integration - To bring ML and deep learning capabilities in
- SciKit Learn library integration
- Developing ML use cases using Meteos.
- Data/batch processing on Hadoop/Spark using OpenStack Sahara project -
No need to put special effort for setting up analytics stack.
- Data is stored and scaled automatically on swift.
*Goal - *
- Users should be able to implement data science problems.
- Data visualization on Portal.
- Developer should be able to develop Machine Learning application
quickly and easily.
- Data storage, Data processing and Data Storage should be achieved
effortlessly and effectively.
If you are interested to work/contribute in this Project, please reach out
to me/hiroyuki on #openstack-meteos or reply back to this mail. My irc
nickname is diga.
OpenStack Development Mailing List (not for usage questions)