ECASLab is a scientific data analytics environment built on top of ECAS (the ENES Climate Analytics Service), one of the thematic services included in the EOSC-hub service portfolio.

It provides a scientific environment exploiting a server-side approach and integrating both data and analysis tools to support data scientists in their daily research activities.

ECASLab starts from a previous effort (OphidiaLab, developed at CMCC Foundation) with the main aim of providing a virtualized research environment for researchers. It represents the entry point for users that want to test, train, exploit the ECAS Thematic Service.

A few examples of output related to different analytics experiments implemented in the ECASLab environment.

It consists of several components like an ECAS cluster, a JupyterHub instance jointly with a large set of pre-installed Python libraries for running data manipulation, analysis, and visualization, a data publication service and a tool for the infrastructure monitoring (mainly intended for the administrators).

In order to get started with ECASLab please have a look at the Quick Start section and register here to get an account.

Quick Start

ECASLab provides two different ways to get access to its scientific eco-system: JupyterHub and Ophidia client.


Jupyter supports interactive data science and scientific computing.
ECASLab includes a JupyterHub installation and, thanks to the Jupyter Notebooks, scientists can create and share documents that contain live code, equations, visualizations and explanatory text.

The JupyterHub interface is available here*.

After you login, open "Quick Start.ipynb" notebook available under the quickstart/ folder in your home to get started with OphidiaLab environment capabilities.

*Please note that for security reasons, the access to our JupyterHub instance is restricted to authorised users only and needs an additional step after the registration process.








The Ophidia Terminal is a robust, comprehensive, and user-friendly Ophidia client, developed with characteristics similar to the bash shell present in almost all Unix-like environments.
Please have a look at the online available documentation to learn more about the basic functionalities of the Ophidia terminal as well as some advanced features useful for more skilled users.
Two short guides (basic, advanced) in pdf format are also available.
Several examples of real-world usage of the terminal are also available on the Ophidia website tutorial section.
The latest client RPM for CentOS7 is available here.
The related DEB package can be downloaded from here.

Once installed you can simply run:

/usr/local/ophidia/oph-terminal/bin/oph_term -H ophidialab.cmcc.it -u <username> -p <password> -P 11732

where <username> and <password> are the ones you will get through the registration process.
A comprehensive user guide about the Ophidia Terminal is available here.

JupyterHub

A simple example about a Jupyter notebook interacting with the Ophidia instance through the PyOphidia Python class.

Experiments

The Experiment section lists a series of example workflows that can be executed within the ECASLab environment to perform different kind of analysis on scientifica data.


Monitoring

The Ophidia framework allows the concurrent execution of single operators, massive tasks and workflows of tasks. In this context, it is critical to monitor the cluster resources usage and activity from both an infrastructural and an application-level point of view. The environment includes an instance of the Grafana monitoring system, which allows the definition of interactive dashboards summarising some relevant metrics, like:

  •  system load average (1/5/15 min.);
  •  ram, swap and disk usage;
  •  workflows progress ratio (percentage of workflow completed);
  •  number of concurrent workflows and tasks in running and pending states;
  •  cores used;
  •  services status.