Excelero delivers big data AI storage solutions for business & enterprise, big data storage solutions, and enterprise data storage solutions. Applications are for major web scale companies for data analytics, machine learning applications in media and entertainment and HPC environments. Skip to main content
ExceleroLatest

The Innovation Milestones Keep on Coming

By July 14, 2020July 27th, 2020No Comments
High performance computing storage

Innovation happens in big and small ways. Big ways, like award-winning products that enable AI/ML/deep learning storage for some of the biggest brands on the planet – delivering results that other solutions just could not. Small ways, like patents that demonstrate our DNA lies in doing things differently and better than the rest.

Excelero quietly achieved another milestone in June 2020 with the award of its fourteenth US patent, our eighth in 2020 thus far. This brings Excelero’s US portfolio to a total of 14 patents, including 10 issued and 4 more allowed with 12 more patents that have been identified or pending with the US Patent and Trademark Office (USPTO).

We are especially proud of US patent 10,740,202 governing a potential improvement for disks (e.g. NVMe devices) that would enable a sizable improement in the implementation of rebuild operations for distributed storage. The improvement would enable skipping areas of the drive that are not currently in use by adding a command to check whether they actually have data on them. Drives know this information, but do not have a means for sharing currently.

On June 28, 2020, we also secured US patent 10,728,331,  2020, governing techniques for dynamic cache use by an I/O device.  NVMesh software can be offloaded to a SmartNIC for two important purposes: freeing the main CPU for other tasks and simplifying deployment models. However, the limited memory bandwidth on the SmartNIC can limit storage offload effectiveness. This patent enables a more elegant way to use host CPU memory to improve SmartNIC performance while still offloading.

Then there’s patent 10,678,455, issued June 9, 2020, which recognizes Excelero’s system and method for increased efficiency of thin provisioning of data storage. NVMesh currently provides state-of-the-art distributed storage with great elasticity, performance and data protection. This patent covers technology to expand NVMesh software capabilities with additional enterprise storage services, specifically distributed thin provisioning, without significantly reducing the performance levels, which is a typical issue with such services.

Just one week earlier, on June 2, 2020, Excelero’s patent 10,673,947 was granted, covering a system and method for providing a client device with seamless access to a plurality of remote storage devices presented as a virtual device. As AI/ML/DL and HCP applications drive astounding growth in data volumes, finding smarter ways to virtualize traditional physical storage is fundamental. With this patent Excelero is setting a new, higher bar for AI/ML/DL storage efficiency as applications scale and deliver both low latency and high throughput.

These patents follow another three patents that were granted earlier in 2020:

On May 26, 2020, Excelero was officially granted patent 10,664,191 for a system and method for providing I/O determinism based on required execution time.

Excelero’s May 12, 2020 award of patent 10,649,950 for disk access recovery techniques, covers an alternative means of improving recovery from errors with NVMe drives in a distributed storage domain. Existing implementations may suffer from correctness issues.

Another Excelero patent from February 18, 2020, patent 10,565,054,  governs a means of reducing the network bandwidth consumed with distributed erasure coding protected volumes for distributed storage. The most novel technology developed as part of NVMesh 2.0 is the ability to provide state-of-the-art performance and data distribution and elasticity while protecting data at the highest levels, using erasure coding. It is the technology embedded in patents such as this that make it possible.

Startups that innovate, succeed. As data center teams re-invent architectures to avoid traditional controller limitations, they need to power their AI/ML/DL and analytics applications with the option to leverage new technologies such as SmartNICs, and smarter approaches to enterprise-caliber data services, virtualization everywhere even with remote devices, superior I/O execution times, highly accurate NVMe disk recovery and ultra-low bandwidth consumption techniques. Excelero’s innovations already are helping these demanding applications to achieve better storage efficiency and performance from shared NVMe Flash resources across an enterprise.

Backed by these most recent 2020 innovations, Excelero is enabling organizations such as Orange, CCC, Barclays, Bezeq, STFC, and more to operate at peak performance and ROI – calling NVMesh “a total game-changer for AI.”

For more details on Excelero’s recent patents, please check United States Patent and Trademark Office patent listings:

PAT. NO. Title
1 10,678,455

System and method for increased efficiency thin provisioning with respect to garbage collection

2 10,673,947

System and method for providing a client device seamless access to a plurality of remote storage devices presented as a virtual device

3 10,664,191

System and method for providing input/output determinism based on required execution time

4 10,649,950

Disk access operation recovery techniques

5 10,565,054

System and method for reducing bandwidth consumption of a client device accessing a remote storage device

6 10,372,374

System and method for providing input/output determinism for servers hosting remotely accessible storages

7 10,237,347

System and method for providing a client device seamless access to a plurality of remote storage devices presented as a virtual device

8 9,971,519

System and method for efficient access for remote storage devices

9 9,658,782

Scalable data using RDMA and MMIO

 

Yaniv Romem

Author Yaniv Romem

More posts by Yaniv Romem