Data Built for AI
Redefine how data is organized, accessed, and protected so enterprise data is ready for foundation model use cases.
Turn enterprise data into a foundational asset for the AI era
Simba Data Foundation
Simba Data Foundation is one of two core pillars of the enterprise AGI platform, alongside the model foundation. It connects enterprise-scale data assets with AI capabilities, bringing together data ingestion, governance, knowledge capture, and secure delivery so models and intelligent business applications can access trusted, governed data on demand.
Let data flow in ways that models can understand, businesses can reuse, and security teams can audit.
The data foundation goes beyond a traditional data platform. It organizes enterprise data into production-ready assets that AI and other IT systems can reliably consume and reuse.
Data Stays in Your Domain
Support full-stack private deployment so enterprise data remains within your environment, helping protect data sovereignty and support compliance.
Knowledge as an Asset
Turn metrics, business rules, industry practices, and expert knowledge into structured assets AI can reuse.
Connect Once, Reuse Everywhere
Unify integration, governance, and delivery to reduce duplicate work and accelerate the journey from AI pilots to scaled deployment.
Store More. Compute Faster. Run Reliably. Govern Clearly. Control Costs.
Keep high-value data ready to support models and business decisions.
500B records/day
Data Throughput
Deliver stable, model-facing throughput for EB- and PB-scale data flows.
1M+ QPS
Query Throughput
Low-latency, high-throughput data APIs through a unified interface, supporting 100K+ QPS.
100B+ events
Real-Time Processing
Unified stream and batch processing, delivering query results within seconds—even for complex queries spanning large datasets, multiple sources, joins, and business rules.
1/3-1/5
Storage Compression
High-efficiency encoding reduces storage requirements to one-third to one-fifth of their original size.
Simba Data Foundation: A Governed Data Hub for AI Models
Bring together enterprise data sources, governance systems, knowledge assets, and access controls to provide AI applications with governed, trusted, on-demand data context.
AI Applications / Agents / Business Scenarios
Unified accessDeciAGIDataAGIOpsAGIOmnichannel AllocationStore IntelligenceSales ForecastProduct RecommendationSupply Chain IntelligenceBusiness DecisionsEnterprise Knowledge Assistant
Unified Data Service Interface
Low Latency, High Throughput, and AuditabilityModel data APIsContext buildingOn-Demand Schema AccessCaching and accelerationPermission checksAudit trails
Multi-Source Data Access
Unify databases, warehouses, data lakes, APIs, files, and streaming data.
Unified Data Management
Identify schemas, field semantics, and relationships while managing catalogs, lineage, and lifecycle.
Data FoundationMetadata Center · Semantic Layer · Metrics System
Enterprise Knowledge Assets
Transform metric definitions, industry rules, business logic, and expert knowledge into structured, reusable AI assets.
Security and Access Governance
Bring together row- and column-level permissions, on-demand context loading, audit trails, and private deployment controls.
Enterprise Multi-Source Data
Unified ingestion · Unified governanceRelational databasesData warehousesData lakesBusiness APIsDocumentsStreaming data
InIngestion
Connect sources in minutes
Connect sources in minutes
GovGovernance
Manage metadata, lineage, and quality
Manage metadata, lineage, and quality
KnKnowledge
Reuse metrics, rules, and expertise
Reuse metrics, rules, and expertise
SecSecurity
Private deployment, permissions, and audits
Private deployment, permissions, and audits
APIDelivery
Reliable model-facing services
Reliable model-facing services
Positioning
Traditional data platforms serve people. This data foundation serves models. It helps models understand business terms, honor access controls, access trusted data, and continuously turn enterprise knowledge into reusable intelligence assets.
Build a secure, trusted data flow from ingestion to model calls.
Simba brings together schemas, metadata, metrics, semantic layers, access policies, and caching to create a model-ready data service layer, so downstream applications receive stable, accurate, and auditable data context.
Multi-Source Data Access
Connect relational databases, data warehouses, data lakes, APIs, files, and streaming data using low-code connectors.
Unified Governance and Knowledge Capture
Ingest schemas, manage metadata and lineage, and capture metrics, semantic layers, industry rules, and expert practices.
Secure Delivery to Models and Agents
Use row- and column-level permissions, on-demand context loading, and audit trails so models only see authorized, business-aligned data.
Simba Data Foundation helps AI understand your business, use real data, and respect defined boundaries—not just answer questions.
General-purpose models are widely available, but AI that understands your business, uses your data, and respects your boundaries must be built on your own enterprise data foundation. Simba turns data assets into a continuously evolving enterprise intelligence layer.
DataBlack: Full-Lifecycle Data Security Engine
DataBlack is StartDT's data security engine. Built on a data-centric security architecture, it helps enterprises protect data across the full lifecycle, strengthen governance and risk management, and safeguard data assets.
Core Technologies
More comprehensive. More intelligent.
Full-Lifecycle Control
Apply comprehensive security controls across data flows and sensitive data.
Intelligent Classification
Apply intelligent algorithms and trained models to identify sensitive data faster and more accurately.
Comprehensive Audit
Provide detailed audit records, including sensitive-data usage, to support security and compliance requirements.
Masking and Encryption
Support static and dynamic masking, encryption, and authorized decryption to protect sensitive data in use, at rest, and in transit.
Risk Detection and Monitoring
Monitor high-risk user actions with custom rules or intelligent algorithms, and trace activity through security audit logs.
Product Architecture
Security Throughout the Data Lifecycle
DataKun: Independently Managed Data Storage and Compute Engine
DataKun is an independently managed storage and compute engine that helps enterprises build a lightweight, intelligent data platform and develop their own analytics and processing capabilities. It supports multiple big-data jobs and services, custom components, validated version combinations, and continuous upgrades of core components.
Product Advantages
Technology You Control. Cost Efficiency.
Ready to Use Out of the Box
- Deploy within an hour and create clusters in minutes.
One-Click Migration
- Compatible with major open-source big data ecosystems.
- Cluster migration tools help customers migrate quickly and smoothly.
Controlled Technology
- Storage layer: supports HDFS and StartDT's SFS, which separates compute from storage.
- Engine layer: supports community-supported open-source and custom components with ongoing updates.
Intelligent Operations
- Layered operations and rich metrics help teams quickly detect and diagnose issues through visual monitoring.
Cost Efficiency
- Separating compute from storage reduces storage costs.
- Operations management tools reduce operating costs and manual effort.
- Lower license costs reduce vendor lock-in risk.
Secure and Reliable
- Supports high availability and high performance.
- Enterprise-grade data security capabilities support security, reliability, and compliance requirements.
Product Architecture
Open, manageable, extensible, and continuously evolving
