Product · Deploy Anywhere
Four deployment models. One codebase.
Cloud, hybrid, on-prem or fully air-gapped, with the same screens, features and security model in each. Accreditation travels with the platform.
Cloud
AWS, Google Cloud or Azure, in your account and region.
Hybrid
Classification decides where each workload runs.
On-prem
The full platform on your own hardware.
Air-gapped
No route out. The design baseline.
Targets and runtimes
Five targets. The same platform on each.
The same platform reaches every target. The air-gapped line is cut on purpose: updates arrive on verified offline media.
One codebase · same screens, features and security model
Amazon Web Services
Your AWS account and region.
Google Cloud
Your project, region and keys.
Microsoft Azure
Your subscription and region.
Your data center
Your hardware and network.
Air-gapped
No route to the internet.
Boundaries and updates
What stays where, and how updates arrive.
- CloudBoundaryYour cloud tenancy, region and encryption keys. Syntasa holds no standing access.Update pathCustomer-approved releases
- HybridBoundaryClassified or regulated processing on-prem; elastic work in your cloud tenancy.Update pathSplit, approved per substrate
- On-premBoundaryYour data center, hardware and network controls, end to end.Update pathOffline media or a controlled channel
- Air-gappedBoundaryNo route to the internet at any time. All dependencies vendored; no callbacks, telemetry or external model APIs.Update pathVerified offline media, customer-controlled promotion
Design baseline
Air gap is where we start, not an edition.
Air-gapped operation is the design baseline. The other three models are relaxations of it. Everything installs from offline bundles and runs with zero external dependencies.
- No callbacks and no telemetry, ever
- No external model APIs; models run inside
- All dependencies vendored in the bundle
- Offline price lists for runtime cost estimates
Accreditations
Runtimes
Run where your compute is.
Choose a runtime per workload. Streaming variants are available where the engine supports them.
Amazon EMR
Managed Spark clusters in your AWS account.
Google Dataproc
Managed Spark on Google Cloud.
Azure HDInsight
Managed Spark on Azure.
Kubernetes
Container runtimes on your clusters.
Docker containers
Your own images as runtimes and steps.
On-prem Spark
Your existing Spark, in your data center.
Available on
Runtime management
Start, stop and size without a ticket.
Runtime templates, schedules, autoscaling, GPU options and cost estimates in one place, with sharing and ownership clear on every runtime.
- Scheduled start and stop, idle shutdown, terminate on completion
- Maintenance-window pause that restores schedules exactly
- GPU runtimes for training and RAPIDS work
Connections and secrets
Your sources, your secret stores.
Secrets stay in AWS Secrets Manager, Google Secret Manager or Azure Key Vault.
Cloud storage
Warehouses & databases
Analytics & marketing feeds
APIs & files
Git
Secret stores
Secure, accredited and observable anywhere.
Map out your deployment.
Tell us where your data must stay. We will walk through the architecture, update path and accreditation for your environment.