Excelero, a disruptor in software-defined block storage, and ThinkParQ announced
that their joint go-to-market relationship is bearing fruit and the
firms have successfully integrated Excelero's NVMesh software-defined
block storage solution for shared NVMe at local latency along with
ThinkParQ's leading parallel cluster file system BeeGFS. Benchmark
results from early deployments spotlight the maximum resource
utilization and flexibility that are possible from combining the two
solutions - performance that is perfectly suited for data-intensive
workloads such as high performance computing (HPC), AI, ML, and
analytics.
In
response to customer demand, Excelero and ThinkParQ already have
deployed NVMe storage clusters with BeeGFS and NVMesh to increase
productivity of advanced GPU compute clusters. These organizations
experienced the storage bottleneck issue endemic to AI - that I/O
intensive applications cannot be processed fast enough even on GPU-based
systems. As a result, IT teams often copy datasets in and out locally,
which is an inefficient and time-consuming process.
"With
Excelero's NVMesh, our customers have access to an ultra-low latency,
high performance approach to scale-out storage," said Frank Herold, CEO
of ThinkParQ. "We've been impressed with NVMesh's ability to deliver the
high IOPS and ultra-low latency of NVMe drives over the network with
highly available volumes - as well as options for distributed erasure
coding and BeeGFS' unmatched ability to efficiently handle all kinds of
access patterns and file sizes."
To
demonstrate the possibilities of this new scale-out infrastructure, the
companies used the industry standard mdtest and IOR benchmarks. The
test system was a compact 2U 4-server chassis with a total of 24 NVMe
drives, connected via a 100Gbit RDMA network to 8 BeeGFS client compute
nodes. Tests were run on the exact same hardware configuration with
BeeGFS utilizing the direct-attached NVMe vs. BeeGFS utilizing NVMesh
logical volumes. Taking advantage of NVMesh to offload mirroring
operations, BeeGFS file create operations were boosted 3x, while
metadata read operations were boosted 2.5x. For small random file
access, which is often considered to be especially critical for
application efficiency, NVMesh's low latency technology boosted BeeGFS
4K write IOPS to 1.25 million per second, a 2.5x improvement. By
leveraging NVMesh distributed erasure coding for BeeGFS, customers can
get up to 90% usable capacity while still tolerating drive failures. All
the while, achieving 75GB/s streaming throughput from this entry-level
system.
Through
June 30 Excelero is offering enhanced pricing on combined purchases of
NVMesh and BeeGFS. More details are available from both companies.
"You
can't build tomorrow's enterprise on yesterday's infrastructure," said
Lior Gal, CEO and co-founder of Excelero. "We're delighted at how
smoothly NVMesh and BeeGFS work together, and the way we enable
organizations to work without compromise to either IT teams or end
users. We look forward to working further with ThinkParQ so that more
organizations worldwide can maximize their NVMe ROI and the efficiency
of their entire operations."