Integrated lithium battery energy storage solutions
Technical Support Downloads
TIANCHU Get a Quote
CLOUD-EDGE ARCHITECTURE | Cloudside Co-ordinated Structure

Edge ensures on-site autonomy,
Cloud support scale operation

Full-stack cloud–edge architecture keeps device control, safety interlocks, data acquisition, and local apps at the site edge; places cross-site analytics, unified O&M, lifecycle management, and scaled services in the cloud.

(b) Once the network is restored, data, configurations and events can be sent and reconciled in a controlled order.

Clouds (Attic Cloud Platform)

Data and analysis
Policies and models
Assets and prosecution
API Services
Security Line / VPN

Regional Operations Centre

Marginal layer (station edge platform)

Data collection
Local application
Cache & Forward
Protocol Conversion
BESS Site (In situ Autonomy)
Battery systems
PCS
Control and protection
Edge Server
Station control.
SITE System Sites 120+ Global operations
EDGE Edge Nodes 600+ Deployed
HA Local autonomy capability ≥99.99% Availability
BUF local cache ≥72h Breaknet Buffer
UPD Controlled update success rate ≥99.5% Nearly 90 days

01 Layer deep: connecting structure from device to cloud

L1 Device Layer Core / Model, BMS, PCS, Protector, Sensor
L2 Protection and control layer BMS / PCS controller, protection relay, fire and safety chain
L3 Edge Layer Industrial edge gateway, edge server, data cache
L4 Site Platform Layer Local SCADA, alarm system, reports, transport tools
L5 Cloud service layer Data lakes, analytical models, asset management, API
L6 User and Operations Level Regional Operations Centre, Remote O&M, third-party system
Device level: Equipment layers retain real on-site objects such as cores, modules, BMS, PCS, protective devices, sensors, etc., which are the physical source of all digital models and control logic.

02 Boundary of responsibility: On-site autonomy, cloud collaboration

Level Principal duties Key competencies Typical response time
Device / Control / Protect Layer On-site real-time control and protection, safety interlocks Hardware closed loop, independent operation <10 ms
Edge Layer Data collection, caches, local applications protocol conversion, rule engine, local storage 10 ms – 1 s
Site Platform Layer Site management, alarms, statements, traffic Local SCADA, traffic flow 1 – 5 s
Cloud service layer Global vision, analysis optimization, strategy down. Big data analysis, AI model, scaleing min
Operational level Remote decision-making, organizational synergy Visualization, scheduling, synergetic processes min
Key principles: (b) The closer to the cloud, the greater the emphasis on cross-site synergy, analysis and scale management.

• Windside synergetic architecture section Noodles. Data & Control Path

Blue Telemetry, Orange Commands, Green State Feedback, Purple Updates use a separate semantic path.
Cloud service layer
Data Lake
Time series database
Analytical engine
Forecast Maintenance
API gateway
Site Platform Layer
Local SCADA
Reports and incidents
Statements and analysis
Operational tools
Configuration Management
Edge Layer
Edge Node A
Main Node
Edge node B
Hot.
Control and protection layer
BMS controller
PCS controller
Protective devices
Monitoring and control devices
Transformer protection
Device Layer
Core / Model
Battery management system
PCS
Electricity distribution / support equipment
Temperature / Fire

04 Autonomy and data are not lost in the offline scene.

01 The WAN link's broken. It's impossible to detect a cloud link.
02 On-site control continues. Control and protection do not depend on clouds
03 Local cache of data UTC Timestamp Continuing Writing
04 Local services available SCADA / HMI continues to work
05 Network Restoration and Verification Verify Time and Data Integrity
06 Orderly Reproduction Reproduction by Time and Priority
07 Cloud-based reconciliation Closing gaps and duplication
Current status: Networks are normal, clouds are running simultaneously.

05 Data pipeline: quality assurance from collection to cloud entry

PRO Protocol collection IEC 61850 / Modbus / CAN / OPC UA
NORM Standard / Cleaning Fields, units and treatment of anomalies
ID Identity and Topography Device ID/ Level Map
BUF local cache Sequence Writing and Breakpoint
CLK Time Sync PTP / NTP Consistency
Q Quality marking GOOD / BAD / SUSPECT
CLD Clouds in library Data Lake / TimeSequence Database

06 Edge application life cycle: controlled release and rollback

01 Structure and Version Runtime / Dependence / Configuration
02 Compatibility check Hardware/ Software/ Protocol
03 Greyscale Preview Small range controlled release
04 Approval Window Personnel / Time / Risk
05 Full link node Batch Updates
06 Health verification Indicator / Log / Link
07 Controlled Extension Zoom in one by one
08 Failed Rollback Restore stabilization version
Structure and Version: Every edge application first declares runtime, dependencies, configuration schema, and target hardware version.

07 High availability and disaster recovery design

Local HA addresses single-point on-site issues and cloud-end HA addresses the continuity of scale services, which cannot be replaced by one another.
PWR Power is high. UPS / Double-Wide DC / Redundant Power Input
NET Networks are highly available Double-link, double-exchange, chain status monitoring
EDGE Edges are highly available Double node master / service automatic takeover
DATA Local data protection SD / RAID, Breakpoint, Validation
CLOUD Cloud height available Multi-available areas, multiple copies and service organization
REC Fault recovery process Detection, isolation, takeover, validation and recovery
Fault detection
Isolate Abnormal Nodes
Switch Master
Validation of local services
Restore cloudside sync
Closed Record

08 Deployment model comparison It depends on the industrial scene.

Dimensions Local private deployment Cloud Trust Mixed Cloud Side
Data attribution All stay here. Integrated cloud management On-site real time + cloud long term
Real time control Local We still need to maintain on-site control. Local control, cloud coordination.
Update Upgrade Client maintenance window Focus Upgrade Harmonized Management + Batch Greyscale
Scope Field engineering team Cloud Operations Team Field + Clouds Convergence
Typical application High-isolation-requirement scenarios Standardized operating scenario Large-scale storage capacity operations

09 Copy from single to full station: Size expansion does not change the security boundary

01 Single-site autonomy Run independently and visualize
02 Multistop access Standardized access and templates
03 Regional Aggregation Unified strategy and insight
04 Internet-wide operations Sizing optimization and synergy
Safety boundaries unchanged: Local controls and protection are always carried out at the edge; clouds are managed, analysed and synergized only through controlled interfaces.

10 Border Commitments

API
API page/ external interface standardization All Northbund interfaces are exposed through controlled API, Schema and version strategies, and external systems are not allowed to bypass the edges and directly operate on-site equipment.
SEC
Network and security commitments All remote connections use identification, minimum access, security tunnels, auditing and failure security strategies.
CTRL
Real-time security control remains on site. The cloud cannot replace local protection, interlocks, or PCS/BMS real-time closed loops.

• Delivery Objects

ARCH Structure and boundary documents Cloud / Edge / Site
RACI Responsibilities matrix Owner / RACI
RUN Offline run programme Autonomy
DATA Data contracts Schema / Quality
APP Apply Template to Margins Runtime / Config
VER Update and Rollback Records Release / Rollback
HA High Available Authentication HA / DR Evidence
OPS Recommendations for deployment and transport Operation Guide