Cloud run by Sagan
Sagan runs all of the infrastructure. You use it and pay for what you use.
- Up and running in days
- Isolated for each company
Sagan Data Platform
One data platform for the whole company
Bring all your data into one place, query billions of rows in under a minute, automate pipelines and work together in secure workspaces.
Big data. Fast runs. Small bills.
Scale
10B rows
A ten-billion-row table still answers. Add machines when needed, release them when done, no re-platforming.
Speed
54×
Faster on 1 billion rows: 42 s instead of 38 min, by splitting the work across many machines.
Cost
65%
Lower processing cost every year: no per-core licences, no idle servers.
01For every role
A workspace for every data team
Data engineers, analysts, data scientists and admins work on one source of data, each within their own access.
Data engineers
Land data, automate pipelines
Connect sources, sync continuously, write and schedule pipelines right in the workspace. Failures are flagged at once.
Data analysts
Fast queries on centralised data
SQL in the browser, on every table you may see. Ask in plain language and get the SQL.
Data scientists
Train models right next to the data
Start a GPU cluster in minutes and read the central data directly, with no copies on personal machines.
Platform admins
Manage people, access and cost
Company sign-in, access down to each table, budgets by department, every query logged.
02Performance
When data outgrows one server
Oracle, PostgreSQL and other transactional databases excel at single transactions. For analysis over billions of rows, one server is no longer enough.
Traditional database, 1 serverSagan Data Platform
54×
faster on 1 billion rows: 42 s against 38 min
6.5 min
for 10B rows, where the single server had not finished after 6 hours
| Workload | Traditional database | Sagan Data Platform | Faster |
|---|---|---|---|
| Monthly revenue roll-up over 1 billion rows | 38 min | 42 s | 54× |
| Join 1 billion transactions with 50 million customers | 2 h 10 min | 3 min 20 s | 39× |
| Load 100 million new rows from business systems | 47 min | 4 min 10 s | 11× |
| Revenue roll-up over 10 billion rows | Not done after 6 h | 6 min 30 s | >55× |
The same retail transaction data and the same queries. The traditional database runs on one server with 32 vCPU and 256 GB RAM; Sagan Data Platform runs on a cluster of 8 machines, each with 8 vCPU and 64 GB RAM.
03Cost
Pay for the hours that work
A database server runs and bills day and night. Sagan compute starts with the work and switches itself off when done, so yearly processing cost is up to 65% lower.
- No per-core licences
- Budgets by department and by project
- Cost reports by hours run and storage
04Centralised data
Every source, one place
From business systems to reports, analysis and machine learning, on one platform with one set of permissions.
05Capabilities
Everything a data team needs, in one place
Continuous data sync
A first full copy, then near real-time changes from databases, files and apps.
Central data, open formats
One home for all company data, in open standards many tools can read.
Compute on demand
Start processing, query or GPU compute in minutes; stop it when the work is done.
Governed SQL
Query in the browser, access down to each table, statements that change data need approval.
Pipeline automation
Pipelines run hourly or daily, monitored, with alerts when something fails.
AI-assisted workspace
A browser workspace wired to your data and compute; an AI assistant writes, explains and fixes code.
AI in every step
- Ask in plain language, get the SQL
- Mapping suggestions when moving data
- Pipelines generated from a description Coming soon
- Pipeline failures explained, with a fix Coming soon
06Security and governance
Open data to the whole company, keep control
Admins see who uses which data, when and at what cost, and can stop trouble before it happens.
- Single sign-onWith company accounts, by role and group
- Access down to each tableEveryone sees only what they are granted
- Risky statements reviewedDeletes and updates need an approver
- Every query loggedWho, when, on which data
- Budgets by departmentWarned before costs run over
- Isolated per companyOwn machines, network and storage
- 09:12an.nguyenRead table salesAllowed
- 09:14minh.tranRead table hrBlocked
- 09:20thu.leDelete 1,204 rows in customersAwaiting approval
- 09:31hoa.phamStart a 4-machine GPU clusterWithin budget
- 09:45an.nguyenRun pipeline revenueAllowed
Monthly budget, Analytics team62%
07Deployment
Runs where your data must live
Two ways to deploy, one platform. Every company gets its own machines and storage.
In your own cloud account
Data stays in your own account and storage. Sagan only operates it.
- Fits banks, telcos and healthcare
- A transparent infrastructure bill, in your name