Scalable High Performance Enterprise File Storage
The shared file system that scales automatically, with unlimited storage and high performance. Never run out of disk again.
14-day trial · no credit card required
In Production · Fortune 500 Since 2013 Finance Media AI / ML Biotech SaaS Research
Throughput per node0.55–72 GB/suncached → cached
Small-file speed vs EFS100×
Durability on S311 nines
Scales out to1000sof servers
Features // Built Into the Client
Scale // Automatic
FTPetabytes of data, thousands of servers, one namespace. Mount on more servers to scale out: nothing to rebalance, nothing to pre-provision.
NODES1→10k CAPACITY→PB REBALANCE0 POOLelastic PROVISIONnone NAMESPACEone
Perf // Local-Disk Speed
FTSpeed like a local drive: log-structured writes, adaptive caching, multithreading. 550+ MB/s read & write per node.
READ552 MB/s WRITE558 MB/s IOlog-struct THREADSmulti FETCHadaptive
Security // End-to-End
FTEnd-to-end encryption with a key only you hold. Data is encrypted on your servers before it leaves, and stays encrypted in transit and at rest.
CIPHERsalsa20 MACpoly1305 SCOPEe2e
Durability // Eleven Nines
FTYour object store is the storage layer: S3-grade durability and availability, with always-on end-to-end integrity checks above it.
DURABILITY11×9s COPIESaz×3 CHECKSUMcrc INTEGRITYalways-on LOSS0
Ops // No Storage Cluster
FTNo metadata servers, no intermediate servers, no single point of failure. The client connects your servers straight to the object store.
META-SERVERS0 INTERMEDIARIES0 SPOF0 PATHclient→object-store UPKEEP0
Compat // Everything You Run
FTA standard POSIX mount on Linux and macOS. Existing applications, tools and pipelines run unmodified: no SDK, no rewrites.
MOUNTfuse APIposix LINUXok MACOSok SDK0 REWRITES0 PIPELINESunmodified
Direct to your object store
The ObjectiveFS client runs on your servers and connects directly to the object store. No hop in between, nothing else to operate.
Cloud, on-premise or both. Switching backends changes nothing in your applications.
Consistent everywhere
Every server sees the same filesystem. Updates on one mount are visible on all mounts, across instances, regions and platforms.
Infrastructure that gets out of your way
Storage is the part of the stack that cannot have a bad day. ObjectiveFS has carried production workloads for over a decade on the most durable storage layer in the industry.
- Durability
- 99.999999999% on S3
- Throughput
- 550+ MB/s per node
- Encryption
- End-to-end · client-side
- Integrity
- Always-on checksums
- Snapshots
- Automatic & checkpoint
- In production
- Since 2013
Faster where it counts, at a fraction of the cost
Benchmarked, published and reproducible: small files, large files, and the full feature comparison.
Real output, straight from the harness
$ ./bench testfile
threads stat ops/s round-trips/s total read GB/s
1 1395960 179345 22.28
2 2496894 398616 37.76
4 3599197 651192 50.88
8 4640089 581727 72.75Metadata is where network filesystems struggle. This run measures
stat(2) calls, forced round-trips (getxattr of an attribute that
doesn't exist, so no cache can answer) and cached 1 MB reads,
scaling from one thread to all eight cores.
BAR·stat ops/s vs run max PEAK·round-trips at 4 threads, stat + read at 8 (both marked)
The benchmark is one small C file: download bench.c.
One Namespace // Every Workload
AI // Machine Learning
Feed training and inference from one shared namespace: datasets, checkpoints and models visible to every node, at the throughput GPUs need. Petabyte-scale data on S3 economics, mounted like a local disk, from workstations to Kubernetes.
Media // Rendering
Render farms, transcode fleets and asset pipelines read and write the same filesystem concurrently. Spin instances up for the job and down when it ships. The storage stays, without a storage cluster's overhead.
Finance // Compliance
Client-side encryption with your keys, immutable and append-only files and directories, point-in-time snapshots and full ACL support: the controls regulated data demands, on infrastructure you control end to end.
Scale-Out // Services
Web fleets, CMS platforms and Kubernetes workloads share file assets across every server and region. Add capacity by adding mounts. Updates propagate to the whole fleet automatically.
Operator Reports // Trusted by the teams who run it
ObjectiveFS has been a fantastic addition to our infrastructure. I don't think I will ever work with S3 without it again. Principal Architect · WeaveUp Inc
Thank you again for your excellent support and innovative product; both make accomplishing a truly scalable WordPress infrastructure on AWS easy. Senior Technical Architect · Fortune 500 company
We tried so many filesystem solutions and ended with ObjectiveFS. It fits our needs in terms of easy, scalable and performance. We also received really great support. Head of Infrastructure · Kaskus
Ownership // Your infrastructure, not ours
Ownership // Your Bucket, Your Account
The data sits in your own object store, in your own cloud account, under your retention rules. We never hold it, and the keys that decrypt it never leave your machines.
Ownership // Nothing of Ours to Operate
No appliance, no metadata tier, no cluster of ours in your environment. The client on your servers is the whole product. It upgrades on your schedule, one machine at a time.
Pricing // Upfront
Plans and prices are on the pricing page, and every plan starts itself. Enterprise is a conversation. Everything else is self-serve.
PLANS·from $99/mo ENTERPRISE·contact us TRIAL·14 days CARD·not required
Get Started // Mount Your First Filesystem
14-day trial · no credit card · running in minutes