Fundamentals 9 min read

Global HPC Storage Market Share Analysis and Future Outlook (2020‑2024)

The article provides a comprehensive analysis of the high‑performance computing (HPC) storage and networking market from 2020 to 2024, detailing component growth rates, vendor market shares, storage media trends, network interconnect preferences, and upcoming exascale system designs across the US, Japan, China, and Europe.

Architects' Tech Alliance
Architects' Tech Alliance
Architects' Tech Alliance
Global HPC Storage Market Share Analysis and Future Outlook (2020‑2024)

From a storage and network segmentation perspective, storage is the most attractive and fastest‑growing HPC component (CAGR 8.4%), accounting for about 20% of total HPC spend; roughly $0.4 of every $1 spent on compute goes to storage.

Hyperion Research released an updated HPC market summary and future forecast report, referencing the 2020 HPC market summary and forecast (download link provided).

In the HPC storage market, Dell EMC, IBM, HPE/Cray and DDN rank as the top four vendors, with Dell EMC leading at 26.3% share and also dominating the academic‑research and industrial sectors; HPE/Cray is third overall but first in the government sector.

Hyperion classifies storage providers into system integrators, system vendors, independent storage vendors, cloud providers, and customer‑built solutions; 73.1% of vendors are system providers (e.g., Dell EMC, IBM, HPE), while DDN is the only independent storage vendor among the top four (10.3% share).

Analysis of parallel file system data over three years shows NFS still holds the highest market share but is rapidly declining, while Lustre, Spectrum Scale/GPFS, and HDFS are growing quickly, with Lustre leading. GFS, Ceph and PanFS have low shares and are grouped as “Other”.

Lustre is most used in academic and government research, whereas industry mainly uses NFS. SSDs are becoming the mainstream storage medium; 45% of capacity in 2019 deployments was SSD, and 86% of sites have flash storage (15% all‑flash, 71% hybrid, 14% all‑HDD).

In network interconnects, InfiniBand remains dominant but Ethernet is growing fast; both are expected to be the top two HPC network choices, while Fibre Channel and local bus technologies continue to shrink.

Future HPC networking must not only provide high bandwidth but also optimize performance for GPU‑accelerated AI workloads, support data movement between on‑premise and cloud, and meet edge‑computing demands.

Major US exascale projects (Aurora, Frontier, El Capitan) each have budgets exceeding $1.8 billion. Aurora plans to integrate Intel’s 7 nm Ponte Vecchio GPU with Xeon CPUs, targeting >1 EF performance, with detailed CPU/GPU specifications listed.

Frontier (CORAL‑2) is the first US exascale system, featuring AMD EPYC CPUs and Radeon Instinct GPUs, Cray Slingshot interconnect, and a liquid‑cooled architecture delivering 40 MW total power.

El Capitan (CORAL‑2) also uses next‑generation AMD EPYC processors and Radeon Instinct GPUs, with similar Cray Slingshot interconnect and a 30‑40 MW power budget.

Japan’s Fugaku system topped the TOP500 in June 2020, achieving 415.5 PFLOPS on the HPL benchmark using Fujitsu A64 ARMv8.2 processors, without GPU acceleration.

China is developing three prototype supercomputers (NUDT, Sugon, Sunway), with at least one expected to enter full production.

The European EuroHPC initiative, launched in 2018, involves 32 EU countries building a 150,200 PFLOPS system (Finland, Spain, Italy) with €650 million investment, plus a €180 million mid‑scale 4 PFLOPS system; further large‑scale procurements and a hybrid HPC/quantum system are planned for 2027.

The article concludes with a list of upcoming report topics, including HPC market summaries, cloud business analysis, storage/network/server analysis, quantum computing scenarios, exascale system analysis, AI market updates, and innovation awards.

storageNetworkingmarket analysisHPCexascale
Architects' Tech Alliance
Written by

Architects' Tech Alliance

Sharing project experiences, insights into cutting-edge architectures, focusing on cloud computing, microservices, big data, hyper-convergence, storage, data protection, artificial intelligence, industry practices and solutions.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.