Content for the HPC community and innovation enthusiasts: tutorials, news and press releases for users, engineers and administrators
- All News
- Aerospace & Defence
- Artificial Intelligence
- Cloud Platform
- Cloud Platform
- E4 Various
- European Projects
- Kubernetes Cluster
- Latest news
Data Science: E4’s range enriched with the announcement of URANIA
E4 Computer Engineering announces a high-performance native cloud solution designed to host distributed Big Data Analytics, Machine Learning and Deep Learning workloads in support of Data Science
E4 Computer Engineering‘s range of solutions is enriched by a new component with the arrival of URANIA, a cloud-native architecture for high-performance and scalable Big Data Analytics, Machine Learning and Deep Learning workloads supporting Data Science activities. URANIA makes management simplicity, an intuitive approach and the flexibility of state-of-the-art systems its strengths.
URANIA is the ideal solution to host the entire life cycle of the latest open-source Generative AI models and can transparently integrate local or remote Large Language Model (LLM) services directly into Jupyter notebook environments to generate new AI code or accelerate the data scientist’s operations.
With its characteristics of great versatility, capacity and computing power, URANIA allows data scientists and data analysis specialists to focus on their discipline without having to deal with the configuration of the operating environments or the harmonisation of the hardware and software components that implement them.
The architecture provides transparent and timely access to all cluster resources, exponentially speeding up user work times, with significant benefits in terms of efficiency and cost reduction.
“In a world increasingly driven by the availability and intelligent use of information, the ability to conduct extensive and in-depth analyses on the data available to companies and organisations can make all the difference, economically but also in terms of operational efficiency,” explains Cosimo Gianfreda, CEO of E4 Computer Engineering. “URANIA opens a new era in data science, providing specialists with unique and advanced functionalities, capable of significantly accelerating any research thanks to the use of an ultra-high performance multi-node environment.“
An end-to-end solution, URANIA is based on KAPTAIN, a Kubernetes cluster that also includes a Web UI, designed and configured to host computationally intensive workloads, and on the E4DS-platform, the cloud-native platform for Data Science, designed and maintained by E4 Analytics, which integrates a series of open-source systems, services and frameworks from the most accredited communities in the sector, to guarantee an infrastructure capable of hosting all stages of the data science workflow.
Data Science is the discipline of transforming data into useful information and, to do so, integrates the principles and practices of computer science, mathematics, statistics and machine learning. In a constantly evolving and increasingly digital world, having valuable information is a priority and knowing how to exploit it is crucial to the success of every organisation and business.
To support this requirement, URANIA includes the main distributed data processing systems and popular open-source frameworks for Machine Learning and Deep Learning, provides services for interactive computing in Jupyter Notebook technology, as well as a cloud native batch scheduler ideal for launching distributed data processing jobs on multiple nodes, in multi-user scenarios. As an open and regularly updated solution, it allows users constant access to the latest technologies
Looking at its components in more detail, the basic version of the Kubernets KAPTAIN cluster is configured with:
- Storage Class by default to guarantee the persistence of data to containers instantiated on the infrastructure.
- NVIDIA’s Kubernetes Operators to provide containers hosted in the infrastructure with the ability to both directly access the GPUs on the Worker Nodes and utilise the capabilities of the RDMA Network.
- Intuitive and modern Web User Interface, which allows the administrator to manage and monitor the system and user activities, and for users to configure and deploy their containerised workloads, within the resource utilisation limits defined by those administering the cluster.
The E4DS-PLATFORM, which provides the interactive environment for Data Science, is equipped with the elements designed to maximise the productivity of Data Scientists and Data Engineers, making the systems, services and frameworks required for the management and implementation of each stage of the typical workflow in the field easily accessible to them.
“Adopting a solution like URANIA can offer numerous benefits, including greater flexibility, scalability, resilience and ease of development and deployment of Data Science applications,” explains Mario Rosati, CEO of E4 Analytics. “It also allows us to make full use of our resources and services, optimising performance and reducing operational costs.“
URANIA further testifies to E4’s constant commitment to innovation and the development of technologically advanced solutions to support companies and expands E4’s already rich and varied offer, which boasts a range of solutions capable of responding to the most technologically advanced needs and covering some of the most innovative areas of the technology landscape.
Find out about the solution here: URANIA