Accelerating Biomedical Research & Discovery
SIB’s state-of-the-art research data infrastructure is underpinned by WEKA® NeuralMesh™ to deliver fast, secure biomedical research outcomes.
0
Scientists in the Swiss Bioinformatics Community
0-Node
GPU-Enabled, Multi-Tenant OpenStack Cluster
0TB
of Object Storage Tiered from NVMe
Introduction
Federating Switzerland’s Bioinformatics Community
Swiss Institute of Bioinformatics (SIB) is an internationally recognized non-profit organization dedicated to biological and biomedical data science contributing to Switzerland’s life sciences and health research. SIB federates the Swiss bioinformatics community of some 800 scientists, encouraging collaboration and knowledge sharing. Its scientists provide essential databases and software platforms, data management, software engineering and biocuration services, IT infrastructure, computational biology know-how, and training. The institute delivers this expertise to academic groups, clinicians, and private companies throughout Switzerland.
The Challenge
Data Shareability and Encryption Without Compromising Performance
SIB provides life scientists, researchers, and clinicians with essential databases, custom tools, computational biology expertise, secure services for sensitive data, and bioinformatics training. SIB also develops BioMedIT: a state-of-the-art data infrastructure that provides a secure, nationwide IT network for advanced biomedical research and informatics.
One of the three nodes of this network, the Romandie Infrastructure at SIB, operates a platform offering biomedical researchers a full service for processing sensitive data. To ensure it could scale to meet demand, SIB needed to find a modern data management platform solution capable of supporting data encryption in-flight, secure shared data access for authorized users only, and exceptionally high performance for applications to support research.
SIB’s IT team evaluated several traditional and new infrastructure alternatives, including BeeGFS and ZFS. None could meet its table stakes requirements for data shareability and encryption without negatively impacting performance. NeuralMesh emerged as the clear frontrunner, providing the performance, security, and sustainability SIB needed to scale its data infrastructure environment.
The Requirements
Performance, Security, and Sustainability at Scale
The new solution needed to:
- Provide advanced data security without impacting performance
- Accelerate application performance to enable faster research and discovery
- Deliver fast, efficient data processing to improve sustainability
- Accommodate mixed data workloads seamlessly
- Offer simplicity and ease of management
The Solution
NeuralMesh Running on Lenovo Thinksystem Servers
SIB selected NeuralMesh with Lenovo servers to provide an accelerated, secure foundation for sensitive biomedical research and data analysis on its BioMedIT infrastructure. NeuralMesh was installed on a GPU-enabled 8-node, multi-tenant OpenStack cluster, which pools virtual resources into a private cloud that can be used to mount it for different tenants and scientific projects. With its advanced data tiering feature, NeuralMesh enables SIB to tier data from its NVMe layer to 100TB of Red Hat Ceph object storage, giving researchers a highly efficient way to access data on demand – for any research project, any time.
We were looking for a solution that is easy to manage, delivers high performance, and protects data with encryption in-flight. This solution ensures secure data access coupled with the necessary data analysis performance required by research groups.
The Outcome
Encryption In-Flight with No Performance Impact
- Increased workload capacity that effortlessly scales with research demands
- Uncompromising performance that consistently fuels research data pipelines
- Multi-tenancy and data access controls for optimal security
- Stellar technical support that minimizes research disruptions
Scale Production AI Faster with NeuralMesh
Your models aren't slow. Your data is. Fix AI bottlenecks with high-throughput infrastructure.