Everything you need to know about storing petabytes of data on-premise in the UK — hardware options, real costs, architecture choices, and how to get started. Fortuna Data has been supplying enterprise storage since 1994.
Petabyte storage refers to the capability of storing one quadrillion bytes of data — equivalent to 1,000 terabytes or one million gigabytes. To put that in context, one petabyte could store approximately 200 million songs, 500 billion pages of text, or 13.3 years of HD video.
Until recently, petabyte-scale storage was the domain of hyperscale cloud providers and the very largest enterprises. That has changed. Falling drive costs, higher-density hardware, and the limitations of cloud storage at scale have made on-premise petabyte storage increasingly accessible to mid-market UK businesses.
Want to know exactly what petabyte storage would cost for your organisation? See our full UK price comparison.
See the cost comparison →Petabyte-scale storage requirements are no longer limited to the largest organisations. Several converging factors are driving mid-market UK businesses toward petabyte storage:
At petabyte scale, the economics of cloud storage break down. Cloud providers charge per gigabyte per month — at petabyte volumes, these costs compound rapidly. Our analysis of Microsoft Azure UK pricing shows that storing one petabyte in the cloud costs over £370,000 across three years. The same capacity on-premise costs approximately £68,396 — a saving of over £300,000.
Beyond cost, on-premise petabyte storage offers advantages that cloud cannot match at this scale: no egress charges when accessing your own data, full UK data sovereignty, predictable performance without shared infrastructure contention, and compliance with UK GDPR data residency requirements.
Thinking about moving petabyte workloads back on-premise? Our cloud repatriation guide covers exactly how to approach it.
Read the cloud repatriation guide →Several hardware approaches are available for organisations looking to store petabytes of data on-premise. The right choice depends on your performance requirements, access patterns, and budget.
The most common approach for active petabyte workloads. Modern arrays like the Seagate Exos CORVAULT provide 1PB+ in 4U with self-healing capabilities and five-nines availability.
Software-defined object storage platforms distribute data across commodity hardware. Ideal for unstructured data at scale — media, backups, archives, and AI datasets.
The most cost-effective medium for cold petabyte storage. LTO-9 tape offers 18TB native capacity per cartridge, air-gap ransomware protection, and a 30-year archive life.
Combines fast NVMe or SSD for active data with high-capacity disk or tape for cold data. Automatically moves data between tiers based on access frequency.
Combines compute, storage, and networking in a single platform. Simplifies management for organisations running petabyte workloads alongside compute-intensive applications.
Separates storage management from hardware, allowing petabyte capacity to be built from commodity components. Reduces vendor lock-in and hardware costs significantly.
At the heart of any petabyte storage system are the storage servers and arrays. Storage servers act as central control units, managing data transfers, access, and retrieval. Storage arrays — which consist of multiple drives working together — provide both the capacity and the data redundancy required at this scale.
Redundant Array of Independent Disks (RAID) configurations are fundamental to petabyte storage resilience. Different RAID levels provide different trade-offs between performance, capacity efficiency, and fault tolerance. At petabyte scale, RAID 6 or erasure coding is typically used to protect against multiple simultaneous drive failures — statistically inevitable when managing hundreds of drives.
Petabyte storage requires high-speed networking to avoid bottlenecks. 10GbE, 25GbE, or InfiniBand connectivity is typically required to achieve the throughput necessary for data-intensive workloads. The network infrastructure is often the component that limits performance more than the storage hardware itself.
Managing petabytes of data requires sophisticated software for tiering, lifecycle management, deduplication, compression, snapshots, and replication. Platforms such as DataCore SANsymphony and IBM Spectrum Scale provide the management layer required to operate petabyte environments efficiently.
Fortuna Data has been supplying and supporting petabyte storage for UK businesses since 1994. Speak to our specialists about the right solution for your environment.
Speak to a specialist →The right architecture for petabyte storage depends on how your data is used. Most organisations operating at this scale adopt a tiered approach — placing frequently accessed data on high-performance disk or flash, less frequently accessed data on high-capacity disk, and cold archive data on tape or object storage.
This tiered architecture delivers the performance of fast storage where it matters, the economics of high-capacity storage for bulk data, and the cost and security advantages of tape for long-term retention — without paying for expensive flash storage for data that is rarely accessed.
Protecting petabytes of data requires a different approach to traditional backup. Full backups of petabyte datasets are impractical — the data volumes are too large to copy entirely on a regular basis. Instead, organisations typically use a combination of:
