Integrating YiTuo ERP with Liucheng Dongli SCM: A Phased Design for 12 Integration Strategies

Scenario and Value

In manufacturing and distribution enterprises, the ERP system is typically responsible for master data and financial accounting, while the supply chain management (SCM) platform handles order flow and upstream–downstream collaboration. Without a stable data channel between these two systems, enterprises often encounter inconsistent product records, misaligned business partner accounts, delayed purchase and sales documents, and unreconciled expense reimbursements. This article focuses on the integration between a YiTuo ERP and a Liucheng Dongli SCM platform, organizing an overall workflow of 12 strategies that span master data, pricing, and business documents. The goal is to help implementation teams clarify execution order, dependencies, and scheduling cadence, thereby reducing joint commissioning and go-live risks.

Integration Architecture and Data Flow

In this integration, YiTuo plays the ERP role, owning the source-of-truth data for products, business partner accounts, warehouses, fund accounts, pricing, and purchase, sales, inventory, and finance documents. Liucheng Dongli acts as the SCM platform, receiving and hosting the corresponding products, business partner accounts, warehouses, fund accounts, pricing, purchase orders and purchase receipts, sales deliveries and sales returns, and employee reimbursement and payment data. The overall data flow is unidirectional from YiTuo to Liucheng Dongli, with no reverse write-back, to avoid conflicts introduced by bidirectional synchronization.

Based on the characteristics of each data object, the 12 strategies are divided into three phases. Phase 1 covers master data and includes four strategies: product synchronization, business partner accounts, warehouses, and fund accounts. Phase 2 covers the pricing system with one strategy. Phase 3 covers business documents and includes six strategies: purchase orders, purchase receipts (with order), open purchase receipts (without order), sales deliveries, sales returns, and employee reimbursement and payment. One additional strategy for inventory requisition is listed as pending configuration because the target interface is not yet ready.

Interface List

Implementation Points

First, there are strict phase dependencies in strategy execution. The four master data strategies in Phase 1 have no mutual dependencies and can be scheduled in parallel. However, the product strategy must execute before the pricing strategy, and the business partner accounts, warehouse, and fund account strategies must execute before the business document strategies for purchase, sales, and employee reimbursement. Otherwise, missing codes or foreign-key reference failures will occur.

Second, writing business documents strongly depends on the code mapping table. Five mapping types must be configured before implementation: material code ↔ product code, business partner code ↔ account code, warehouse code ↔ warehouse code, YiTuo unit ↔ Liucheng Dongli unit, and account code ↔ fund account code. The unit mapping is purely configuration-maintained, while the other four require initial synchronization via the corresponding master data strategies followed by manual configuration.

Third, scheduling frequency should be tiered by data change density. Master data and pricing are recommended to synchronize every 2 hours, staggered by 30 minutes during business hours (07:00–22:00). Purchase and sales documents are recommended to synchronize every 15 minutes to cover business peaks. Employee reimbursement and payment are recommended to synchronize every 30 minutes. Full synchronization should be performed during initial deployment or data repair, while daily operations should rely primarily on increments filtered by updateTime/createTime.

Fourth, exception handling should be tiered. Transient network issues use exponential backoff retry (30s/60s/120s, up to 3 attempts). Target system 5xx errors retry 2 times with a 60s interval. 429 throttling retries 3 times with 60s/120s/300s intervals. 4xx business validation failures and missing code mappings should not retry; they should be logged and skipped with alerts to avoid amplifying load with invalid requests.

Fifth, monitoring and alerting thresholds should be preconfigured. Log single failures. Pause scheduling when consecutive failures reach ≥ 3 and require manual intervention. Trigger alerts when failure rate exceeds 10%, synchronization latency exceeds 30 minutes, missing code mappings exceed 10 entries in a single run, or dead-letter queue backlog exceeds 100 entries.

Best Practices

Perform a weekly full synchronization check for master data every Sunday at 02:00, which covers any incremental omissions without affecting business hours. When the four Phase 1 strategies run in parallel and share the same API throttling quota, staggering or brief serialization is recommended. The inventory requisition strategy currently depends on the /api/qch/GetOrderList API to query source data, while the target-side /dealer/save interface is not yet ready. It is recommended to disable this strategy before the interface is delivered, to prevent dirty data from flowing into Liucheng Dongli. After all 12 strategies are live, continue to observe against the above thresholds for 1–2 weeks, and gradually extend the synchronization frequency of master data and pricing from every 2 hours to longer intervals to save API call costs.