Copyright © 2023. All rights reserved.
cunoFS bridges the gap between your filesystem and cloud object storage
Technology and Benchmarks
Object storage made simple. Objects as files. Files as objects.
Run unmodified POSIX-compliant applications directly on object storage.
cunoFS in action
By unlocking the raw power and scalability of the underlying object storage, cunoFS scales to supercomputing levels of throughput, benchmarked at over 10 Tbps of aggregate throughput.
With competing solutions, handling lots of small files can be impractically slow. In contrast, cunoFS can achieve high performance even on lots of small files. The Linux cp command on 75,000 small Linux Kernel source code files to S3 are reduced from hours in some cases to just over 2 minutes with cunoFS. Reads are reduced from many minutes on EFS, EBS and FSx Lustre, to just seconds.
Unlock the true potential of object storage for your applications by delivering incredible throughputs. Here we can see cunoFS with the cp command is even faster than AWS’ own native CLI tools for transferring data.
Don’t believe our benchmarks? Our benchmark code is public – try it out for yourself. DELL Technologies independently validated cunoFS performance on their own S3-Compatible Object Storage – see their White Paper here.
SPEED
cunoFS
Professional
Professional
cunoFS
Enterprise
Enterprise
EFS
EBS
(gp3)
FSx Lustre
(1.2TiB scratch)
s3ql
goofys
s3fs
Throughput write (cp 5x32GiB files)
50+ Gbps
0.86 Gbps
2.6 Gbps
1.2 Gbps
2.14 Gbps
3.83 Gbps
0.25 Gbps
Throughput read (cp 5x32GiB files)
50+ Gbps
2.1 Gbps
2.8 Gbps
2.1 Gbps
1.45 Gbps
6.62 Gbps
0.75 Gbps
Read Linux kernel source from storage
21 sec
10 min 31 sec
8 min 29 sec
6 min 13 sec
16 min 38 sec
3 min 8 sec
13 min 40 sec
Write Linux kernel source files to storage
2 min 8 sec
5 min 45 sec
16 sec
4 min 5 sec
26 min
1 hour 20 min
2 hours 11 min
COMPATIBILITY
Access objects as files
✔
-
-
Import from S3
-
✔
✔
Access files as objects
✔
-
-
Export to S3
-
✔
✔
POSIX consistency domain
per user
per node
per node
cross user
cross node
cross node
cross user
cross node
cross node
single mount
single node
single node
cross user
cross node
cross node
single mount
single node
single node
per mount
per node
per node
per mount
per node
per node
Atomic rename & hard links
✔
✔
✔
✔
✔
-
-
Mode, UID/GID/permissions, timestamps
✔
✔
✔
✔
✔
-
✔
POSIX ACL
✔
✔
✔
✔
✔
-
✔
Random writes, symlinks, fsync
✔
✔
✔
✔
✔
-
✔
File locking
per user
per node
per node
cross user
cross node
cross node
cross user
cross node
cross node
single node
cross user
cross node
cross node
-
-
-
POSIX permissions enforced at Object API
S3
-
-
-
-
-
-
SCALABILITY
Performance scales with nodes
✔
-
-
by volume size
-
✔
✔
Multi-node support
weak consistency
✔
✔
✔
✔
-
weak consistency
weak consistency
