Software-defined block storage provider Excelero scored a deal with Israeli-based telecommunications provider Bezeq to replace its all-flash array with Excelero’s NVMesh technology to scale-out storage architecture in its mission-critical data warehouse.

Bezeq currently stores more than 118 terabytes of converged data from sales, services, financial transactions, engineering devices, and customer interactions in its data warehouse. The move to Excelero’s platform will allow it to share that information across any network with support for local and distributed file systems. The carrier is also expanding into high-performance computing (HPC) and hosted cloud services.

The deployment uses Fujitsu servers and Mellanox 100G Ethernet switches. All-flash storage typically reduces costs because it is faster and allows companies to shrink their storage footprint. Bezeq said it expects the move to cut database run times by up to 90% and increase throughput by up to 300%.

“The Excelero NVMesh software-defined technology is the only technological solution that meets our performance demands and has proven to be an excellent decision,” said Muginstein Igal, storage and backup team manager at Bezeq, in a press release.

The operator previously migrated from an Oracle Exadata system to a Hitachi UCP with internal Fusion IO and a Hitachi VSP F400 all-flash array connected through a fiber channel. However, it has found that system provided insufficient throughput, slow response times, and lacked expansion capabilities.

More Speed and More Space

The move highlights advantages that NVMe over Fabrics architectures deliver compared to legacy systems. This allows enterprises to run web-scale applications and access data in near real time.

Other vendors in the space include NetApp, IBM, and Dell EMC.

“Traditional storage is dramatically slower,” explained Eric Herzog, vice president of product marketing and management for IBM Storage Systems, as part of that company’s launch of a new all-flash storage array product last year. “What do you do to overcome latency? You have to buy more servers. The value of flash arrays is not only much better application performance in the real world, but it also shrinks your server farms, so you save on capex and opex.”