This topic describes how to use acceleration endpoints of Object Storage Service (OSS) together with Function Compute and CloudFlow to quickly build a cost-effective cross-border object synchronization and transmission system. If your workload is latency-tolerant and prioritizes cost control, you can build this pipeline with no pre-provisioned infrastructure and automatic scaling.
Use cases
This solution fits when all of the following apply:
Cost control: Cross-border data transfer costs must be strictly managed, and dedicated connectivity options exceed your budget.
Latency tolerance: Your application is not latency-sensitive and can accept the synchronization delay introduced by network jitter.
Elastic scale: The data pipeline must handle large-scale file writes without pre-provisioned fixed infrastructure.
Cross-border object transfer: Objects must move between OSS buckets in different regions across geographic boundaries.
How it works
The solution connects an OSS upload event in China (Shanghai) to an automatic object copy in US (Silicon Valley), triggered through Function Compute and a CloudFlow workflow:
An Elastic Compute Service (ECS) instance simulates application writes, accessing OSS buckets in both the China (Shanghai) and US (Silicon Valley) regions to produce and upload data.
When an object is uploaded, OSS triggers Function Compute, which starts a CloudFlow workflow for that object.
The CloudFlow workflow uploads the object to the OSS bucket in US (Silicon Valley) through the OSS acceleration endpoint — one workflow per object, keeping synchronization tasks isolated and independently retryable.
An ECS instance in US (Silicon Valley) reads the replicated objects from OSS to verify data integrity.
Benefits
Low O&M cost: Function Compute handles invocation automatically on each upload event. Focus on business logic, not infrastructure management.
Low network cost: Using OSS acceleration endpoints for cross-border transfer reduces network cost compared with Cloud Enterprise Network (CEN) or Express Connect.
Low deployment cost: No ECS resources need to be pre-provisioned for the sync pipeline. Function Compute invokes CloudFlow on demand, triggered by each object upload.
High elasticity and efficiency: Each object upload triggers an independent CloudFlow workflow. Resources scale automatically with write volume and data synchronization can be efficiently performed.
Solution details
For more information, see Low-cost cross-border file transfer best practices.