
A single distributed infrastructure
Apollo is ICAR’s distributed computing system. It integrates hardware resources, software and expertise available at the Naples, Rende and Palermo sites, making them available as components of a single scalable, multiservice infrastructure.
The resources developed through the FOSSR, H2IOSC and SoBigData projects retain their specific technological and operational characteristics while contributing to the development of shared services for high-performance computing, artificial intelligence, data management and research environments.

Distributed infrastructure and data centre facilities
A new Tier-2 data centre facility has been established at the Naples site. It features hot-aisle and cold-aisle containment to improve the thermal management of the systems. The facility also includes a new fire suppression system, electrical continuity systems and dedicated security mechanisms.

The Rende and Palermo sites also implement solutions for security, thermal management and energy efficiency, tailored to the characteristics of their respective infrastructures.

Resources integrated into Apollo
FOSSR
- 100 Dell PowerEdge R7525 general-purpose servers
- 10 GPU nodes
- 24 NVIDIA GPUs with 80 GB of memory
- Approximately 400 TB of BeeGFS and block storage
- Approximately 1.2 PB of object storage
H2IOSC
- 3 NVIDIA DGX-A100 systems
- 24 NVIDIA A100 GPUs
- 8 TB of first-tier storage
- 129 TB of RAID 1 storage
- 5 Dell PowerEdge R760 servers
- 2 JBOD units
- Kubernetes with 3 master nodes
SoBigData
- 1 Dell PowerEdge XE9680 AI server with 4 NVIDIA H100 GPUs, each with 94 GB of memory
- 2 CPUs, 2 TB of RAM and 20 TB of SAS SSD storage
- 5 Dell PowerEdge R7525 GPU servers
- 8 NVIDIA L40 GPUs, each with 48 GB of memory
- 22 Dell PowerEdge R7525 NVMe servers
- Network connectivity up to 300 Gb/s
- SIPER, Python, Java, Julia and OpenStack

