PEAK:AIO AI Data Servers
Blistering performance for AI workflows, leaving more budget for GPUs
AI Storage by PEAK:AIO
Today's AI servers consume and analyse data at much higher rates than many traditional storage solutions can deliver, resulting in low GPU utilisation and reduced ROI. To gain the performance needed to keep the GPUs busy and fed with data, users often feel the need to choose between feature-rich enterprise-class storage and parallel filesystems designed for HPC.
Not anymore - PEAK:AIO has developed a software-defined storage (SDS) platform from the ground up to complement the NVIDIA DGX, HGX, MGX and EGX range of AI appliances. PEAK:AIO delivers linear scaling, ultra-low latency and huge bandwidth at a price which allows more investment to be made on GPU resource and less on legacy storage - ensuring your project's funds are better spent and your team is more productive.
PEAK:AIO Software-Defined Storage Solutions
PEAK:AIO has three different SDS solutions for different AI use cases - AI Data Server, Open pNFS and Archive, as detailed below:
| Feature | AI Data Server | Open pNFS | Archive |
|---|---|---|---|
| Primary Purpose | High-performance primary storage for AI/ML workloads | Scalable parallel file system and metadata layer for AI/HPC environments | Immutable archive, compliance, and long-term data retention |
| Target Workloads | AI training, inference, analytics, GPU clusters | Large-scale AI, HPC, and distributed NFS environments | AI dataset preservation, compliance, audit trails, ransomware protection |
| Role in Data Lifecycle | Active production data storage | Scale-out data access and metadata orchestration | Long-term retention and regulatory preservation |
| Scalability | ≤ 10 nodes | ≥ 5 nodes | Multiple nodes |
| Capacity | ≥ 100TB | ≥ 1PB | ≥ 1PB |
| Key Benefits | Uses open standard Linux, maximises GPU utilisation, reduces AI storage costs, simple deployment | Uses open standard Linux, eliminates metadata bottlenecks, single namespace at scale, enables parallel data access at scale | Uses open standard Linux, compliance, audit-ready, ransomware protection, immutable storage, instant data recovery |
| Typical Buyer | AI infrastructure teams running GPU clusters | HPC centers, AI platforms, large-scale storage architects | Governance, compliance, security, and AI operations teams |
| Cost | £ | ££££ | ££ |
Explore PEAK:AIO Storage
Discover more about each solution by clicking the tabs below.
PEAK:AIO AI Data Server
PEAK:AIO AI Data Server is a software-defined storage platform purpose-built for AI workloads. Unlike traditional enterprise storage systems that were designed for general IT applications, PEAK AI Data Server is optimised for technologies such as GPUDirect, RDMA, and NVMe over Fabrics (NVMe-oF) to accelerate data throughput directly to GPU resources, delivering up to 160GB/s per 2U appliance. PEAK dispenses with the features that slow down legacy storage solutions, but retains RAID protection for your data, and utilises a traditional NFS file system, typically scaling from hundreds of TB to several PB, depending on the hardware configuration.
PEAK Open pNFS
PEAK Open pNFS is an open standard, vendor-agnostic parallel Network File System (pNFS) platform designed to eliminate storage bottlenecks in AI and HPC environments. It leverages technologies including RDMA, NVMe, and CXL to provide a standards-based alternative to traditional proprietary parallel file systems while enabling storage performance to scale linearly as AI workloads grow.
The pNFS architecture separates metadata services from the metadata store, allowing resources to scale dynamically and independently of the underlying storage, enabling thousands of servers to serve the largest AI and HPC environments.
PEAK Archive
PEAK Archive is a high-density, all-flash archiving and compliance platform designed specifically for AI and machine learning environments. It provides immutable data protection, automated archiving, ransomware resilience, and audit-ready data retention, ensuring that AI datasets, models, and outputs can be preserved, traced and accessed throughout their lifecycle.
As AI moves from experimentation into production, organisations face increasing requirements around explainability, data provenance, and regulatory compliance. The PEAK Archive's high-density, QLC all-flash design delivers multiple PB of storage, so archived data can be presented and accessed immediately without lengthy restore operations.
Ready to Deploy Your AI Infrastructure?
As a leading NVIDIA Elite Partner and the UK's only NVIDIA DGX Managed Services Partner, Scan is ideally suited to design and deploy your GPU-accelerated compute & PEAK:AIO-based storage cluster. Contact our AI experts for further information on 01204 474210 or on [email protected].