IBM Storage Scale System 6000, 340GB/s 4.7 PB 4U Rackmount Storage Array

When you need a storage solution that can keep up with massive data workloads while protecting every bit of information, the IBM Storage Scale System delivers exactly that. It is an active-active dual-controller storage platform built for high-performance environments, offering massive capacity, fast access, and built-in resilience. Whether you are running AI analytics, large-scale cloud services, or demanding enterprise applications, this system removes bottlenecks and safeguards data so you can focus on results, not storage problems.

IBM Storage Scale System - Enterprise-Grade Data Engine

Large data sets can slow down operations, increase costs, and expose you to risk if the storage layer cannot keep up. In today’s fast-moving businesses, slow I/O or data loss translates directly into lost revenue and frustrated users. The IBM Storage Scale System solves these issues with a combination of powerful CPUs, massive memory, and flexible RAID that protects against failures while delivering consistent performance. By integrating advanced data services and supporting a wide range of protocols, it gives you a single, reliable platform that scales with your growth and simplifies management.

Benefits

The system’s dual-socket AMD EPYC™ Genoa processors provide 48 cores, delivering the compute power needed for intensive workloads without compromising latency. With up to 1.5 TB of memory per canister, the platform can hold large working sets in fast RAM, accelerating analytics and AI tasks. Its de-clustered RAID supports multiple erasure-coding schemes, ensuring data remains available even if several drives fail. The built-in Data Acceleration Tier uses NVMe-over-Fabric to boost AI and analytics performance, while the integrated lifecycle management moves data automatically between flash, HDD, cloud, and tape for optimal cost efficiency. Security is baked in with FIPS-140-2 validated 256-bit encryption and IBM Safeguarded Copy, protecting sensitive information at rest and in motion. Finally, the system’s support for POSIX, NFS v4, SMB, S3, and other protocols means you can connect any application without extra adapters.

Usages

• High-performance AI model training and inference
• Large-scale analytics and data-lake operations
• Cloud-native container workloads with Kubernetes CSI integration
• Enterprise backup and archival with automated tiering
• Multi-site disaster recovery using synchronous and asynchronous replication

Key Features

System: Active-active dual controller storage system - continuous availability with no single point of failure.

CPU: Dual socket AMD EPYC™ Genoa CPUs (48 cores) - massive processing power for demanding workloads.

Memory: 768 GB to 1 536 GB per canister (total 1 536 GB or 3 072 GB) - large in-memory capacity for fast data access.

RAID: De-clustered RAID with 3-way/4-way replication, 4+2P, 4+3P, 8+2P, 8+3P - flexible data protection options.

Raw Capacity: Up to 2.2 PB raw per 4U node - scale storage to meet petabyte-scale needs.

Effective Capacity: Up to 5.4 PBe effective - high usable space after data reduction.

Software: IBM Storage Scale for Storage Scale System - integrated management and data services.

Operating System: Red Hat® Enterprise Linux (RHEL) - stable, enterprise-grade OS foundation.

Data Access Services: POSIX, NVIDIA GPUDirect, CNSA, CSI, HDFS, NFS v4, SMB, HTTP, S3 - universal connectivity.

Data Management Services: Lifecycle optimization across memory, NVMe flash, HDD, cloud, tape - cost-effective data placement.

Data Resilience Services: FIPS 140-2, 256-bit encryption, cyberprotect, IBM Safeguarded Copy - comprehensive security.

Data Acceleration Tier (DAT): NVMe-oF performance boost for AI and analytics - lower latency for compute-intensive jobs.

Sequential Read Performance: Up to 340 GB/s - rapid large-file reads.

IOPS: Up to 28 million IOPS (NVMe-oF configurations) - exceptional random-access capability.

Networking / Adapters: NDR / HDR switch support - high-speed fabric compatibility.

PCIe Slots: 16 × PCIe Gen5 slots per system (8 per canister) - ample expansion for accelerators.

SAS Adapters: Up to 12 × 24G SAS adapters (6 per canister) - flexible connectivity for legacy storage.

NVIDIA CX7 Supported Adapters: 400 Gb single-port InfiniBand x16 Gen5; 200 Gb VPI dual-port InfiniBand/Ethernet x16 Gen5 - ready for GPU-driven workloads.

High Speed Network Adapters: Up to 8 CX7 adapters - robust network throughput.

Network Ports: Up to 16 ports - extensive networking options.

NVIDIA Platform Certification: Certified for NVIDIA BasePOD, SuperPOD, Grace Blackwell - seamless integration with NVIDIA AI platforms.

Specifications