Home News

company news about DDN gets Infinia ready for production AI inferencing

All Products

Rack Storage Server
(179)

Huawei Fusion Server
(31)

Dell Poweredge Server
(59)

H3C Server
(31)

Datacom Switches
(96)

WLAN Device
(21)

Smart Wireless Router
(17)

Hard Drive HDD
(78)

Internal Hard Drive SSD
(16)

Geforce Graphic Card
(27)

INTEL CPU Processor
(20)

Server Memory RAM
(6)

Refurbished Storage Server
(6)

SFP Transceiver Module
(4)

Fibre Channel Switch
(125)

Certification

Customer Reviews

The sales staff of Beijing Qianxing Jietong Technology Co.,Ltd are very professional and patient. They can provide quotations quickly. The quality and packaging of the products are also very good. Our cooperation is very smooth.

—— 《Festfing DV》LLC

When I was looking for intel CPU and Toshiba SSD urgently, Sandy from Beijing Qianxing Jietong Technology Co., Ltd gave me a lot of help and got me the products I needed quickly. I really appreciate her.

—— Kitty Yen

Sandy of Beijing Qianxing Jietong Technology Co.,Ltd is a very careful salesman, who can remind me of configuration errors in time when I buy a server. The engineers are also very professional and can quickly complete the testing process.

—— Strelkin Mikhail Vladimirovich

We are very happy with our experience working with Beijing Qianxing Jietong. The product quality is excellent, and delivery is always on time. Their sales team is professional, patient, and very helpful with all our questions. We truly appreciate their support and look forward to a long-term partnership. Highly recommended!

—— Ahmad Navid

Quality： “Great experience with my supplier. The MikroTik RB3011 was already used, but it was in very good condition and everything works perfectly. Communication was fast and smooth, and all my concerns were addressed quickly. Very reliable supplier—highly recommended.”

—— Geran Colesio

I'm Online Chat Now

Company News

DDN gets Infinia ready for production AI inferencing

DDN has rolled out version 2.4 of its Infinia object storage software, stressing this is far more than a routine maintenance patch. The release delivers strategic upgrades tailored for large-scale AI inference, retrieval-augmented generation (RAG), smart data lakes, and multi-protocol workloads catering to research, life sciences, financial services and manufacturing sectors.

The firm points out that as businesses roll out agent-based AI, RAG tools, AI copilots and self-operating AI platforms, inference costs now represent the largest ongoing expense within modern AI stacks. DDN states Infinia’s underlying architecture cuts operational overhead for production inference via sub-millisecond data retrieval speeds, high-throughput object storage and smart data orchestration services, ensuring GPUs stay fully utilized instead of idling while waiting for storage access.

The 2.4 Infinia release introduces an expanded feature suite to simplify adoption for NVIDIA cloud collaborators, managed AI service operators and enterprise AI adopters. The single unified platform can serve distinct teams, external clients, internal departments and isolated sovereign AI tasks concurrently.

Alex Bouzari

Alex Bouzari, DDN’s Chief Executive Officer and co-founder, commented: “AI cost efficiency is quickly overtaking model capability as the top priority for businesses. The industry now judges success by cost per token, inference throughput, GPU utilization rates and tangible business value, rather than just the volume of deployed GPUs. Enterprises need infrastructure that turns costly AI hardware investments into efficient, productive AI operations. Infinia 2.4 delivers the governance, security, performance and operational backbone needed to lift AI return on investment, while laying groundwork for next-gen corporate and sovereign AI deployments.”

New capabilities added in this update include:

-Refined multi-tenant functionality

-Identity linking and centralized access administration
-Rigorous quota limits and governance rules
-Strengthened operational separation for shared AI environments
-Full interoperability with standard S3 deployments and developer SDKs
-Early POSIX compatibility, with validated POSIX clients for RHEL and Ubuntu, fixed throughput targets, plus official deployment guides and operational configuration documentation.

The upgraded multi-tenancy stack allows users to spin up tenants and sub-tenants in roughly 10 seconds through REST API requests. All data objects carry unique tenant identifiers to maintain strict isolation across the whole system, supporting tiered structures (tenant → sub-tenant → dataset). This design enables flexible resource sharing across internal teams, external clients, departments and sovereign workloads hosted on a single cluster.

Sanjay Jagad, DDN Vice President of Product Management, explained the value of POSIX support in a company blog post: “Genomic analysis pipelines, computational fluid dynamics simulations, quantitative financial modeling workflows and digital twin platforms all rely on standard POSIX file behaviors. Previously, running these alongside AI training and inference workloads required separate infrastructure silos, disjointed management processes and redundant data duplication between systems.”

“Now that native POSIX file access is undergoing customer testing, Infinia can accommodate these workloads without requiring standalone storage stacks. File-dependent applications operate on the same platform and identical datasets, retaining the same performance levels and isolation safeguards built for Infinia’s original AI-focused object workloads.”

Jagad further noted that Infinia’s unified architecture enables simultaneous support for inference workloads, intelligent data lakes and cross-protocol tasks without building separate product variants for each use case.

“It is built on a distributed log-structured key-value engine that optimizes every input/output operation individually. Metadata and raw data reside within the same engine, which keeps the data catalog consistently updated in real time. Layout indirection technology supports cluster expansion with zero service interruption, with no data relocation required. Dynamic per-I/O erasure coding delivers tailored data protection for every write operation, eliminating performance drawbacks from rigid stripe sizes. Its tiered keyspace multi-tenancy layer provisions isolated tenant environments within seconds.”

He added: “A single Infinia cluster can host AI training jobs, inference tasks, RAG lookup workflows and POSIX file workloads, with strict logical separation between independent tenants.”

DDN has outlined the following future roadmap for Infinia:

-Large-scale verification: 500–600 node clusters with published linear performance benchmarks
-POSIX general availability: Production-ready functionality integrated with GPUDirect Storage, POSIX access control lists, and native S3/POSIX interoperability on shared datasets
-Enterprise-grade security: Role-based access control integrated with AD/LDAP, dedicated tenant encryption keys via KMS/KMIP, plus WORM and Object Lock functionality
-Data resilience tools: Built-in replication engine and data mover utilities for business continuity
-NVIDIA official validation: Optimized compatibility for DGX hardware plus a formal certification pathway

Full details of DDN’s complete NVIDIA integration roadmap are accessible to qualified partners and enterprise customers under non-disclosure agreements.

Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!

Pub Time : 2026-07-09 11:46:17 >> News list

Contact Details

Beijing Qianxing Jietong Technology Co., Ltd.

Contact Person: Ms. Sandy Yang

Tel: 13426366826

company news about DDN gets Infinia ready for production AI inferencing

Rack Storage Server

Huawei Fusion Server

Dell Poweredge Server

H3C Server

Datacom Switches

WLAN Device

Smart Wireless Router

Hard Drive HDD

Internal Hard Drive SSD

Geforce Graphic Card

INTEL CPU Processor

Server Memory RAM

Refurbished Storage Server

SFP Transceiver Module

Fibre Channel Switch

Rack Storage Server

12 Bays 1U Rackmount Server Lenovo ThinkSystem SR630 Rack Server

ThinkSystem SR250 V2 4SFF Rack Storage Server Intel Xeon E-2378G Processor

Intel C621A Rack Storage Server Inspur NF5180M6 1U Rack Mount Server

Huawei Fusion Server

FusionServer 5288 V6 4U Rack Server 32 DDR4 DIMMs 44 3.5 Inch Hard Disks

Ultra High Density Huawei Fusion Server 1U Network Storage Server 1288H V5

New Gen OceanStor 5310 Huawei Rack Server Hybrid Flash Storage