Real-time video analytics.Process streams near the source.
Real-time video analytics runs on a single-tenant edge cluster close to where your camera and sensor streams start. Process footage near the source and send results upstream instead of raw data.
Analyze video where it lands.Results travel. Raw footage stays put.
Real-time video analytics works best when the processing sits close to the cameras and sensors. Shipping every raw stream to a distant cloud region costs time and bandwidth before any analysis begins.
An edge cluster in a nearby metro data center or your colocation space reads streams over a short path, runs the analysis on local compute, and sends only the results upstream to the systems that need them.
- Shorter path from camera to compute
- Less raw data sent across long networks
- Recordings kept on storage you control
- No egress fees on data leaving the cluster
How the data flows.From stream to result.
- IngestCamera and sensor streams arrive at the nearest edge site over a short network path.
- StoreFootage lands on pooled Ceph storage with NVMe for hot data and HDD for capacity.
- AnalyzeCompute nodes, plus GPU nodes where needed, run your detection and scoring models on the local copy.
- Send resultsOnly events, counts and alerts travel upstream to your central apps or cloud.
Built for steady streams.Node roles sized to your feeds.
GPU and high compute
GPU / high-compute nodes for vision models, alongside compute nodes for the rest of your pipeline.
Tiered storage
HDD, SSD and NVMe tiers pooled by Ceph, with three copies of every block on different nodes.
25/100 GbE switching
Top-of-rack switches connect every node at 25/100 GbE on your own private switching.
Grow by node
Add cameras, add nodes. Each node adds processors, memory and disks together.
Costs that don't grow with every frame.One monthly price for the rack.
Video is heavy, and per-gigabyte pricing turns more cameras into bigger, less predictable bills. An edge cluster runs on a fixed monthly price per rack build, quoted after a free design session.
There are no egress fees and no per-request storage charges. When you add cameras, you add nodes on a plan you already know, and your monthly price changes only when the rack does.
Questions, answered.
Do you supply the cameras or video software?
No. We design, build and operate the cluster that runs your analytics stack. You bring the cameras and the software.
Where can the video analytics cluster run?
In our metro data centers in Chicago, Ashburn or Los Angeles, or in a colocation facility you choose.
What happens if a storage node fails?
Ceph keeps copies of every block on different nodes and rebuilds them on its own, so footage stays available.
Can the same cluster run other workloads?
Yes. The same hyperconverged nodes can run dashboards, databases and the apps that consume your analytics results, on private VLANs.
Keep exploring.
Talk to the engineers who will build it.
Tell us what you run today and where you’re headed. We’ll come back with a design and a fixed monthly price.