Software & SaaS / Vertical ERPBeeda2026-07-10

Beeda: One Serverless Instance per Tenant, Orders and Reporting in One Engine

The wallpaper-industry ERP ran on SQL Server instances each hosting hundreds of customers with roughly 500 tables, where reporting interfered with order entry. Moving to one serverless instance per customer gave every tenant full isolation, with per-SQL billing matching a business that peaks about six hours a day.

Beeda

Beeda is China's specialist software provider for the wallpaper industry, digitalising manufacturer and dealer operations through its own ERP and dealer-ordering product, which encodes the operating practice of dozens of established wallpaper factories and serves thousands of dealers.

1 per tenant
Full isolation
~500 tables
Per customer database
6 hrs/day
Business peak window
Per-SQL
Billing, not fixed capacity

The challenge: hundreds of customers per instance, reports blocking orders

Beeda is China's specialist software provider for the wallpaper industry, encoding the operating practice of dozens of established wallpaper factories into its own ERP and dealer-ordering product, serving thousands of dealers.

The ERP ran on multiple SQL Server instances on Alibaba Cloud, each hosting hundreds of customers with roughly 500 tables per database. One customer running a complex report degraded order entry and inventory lookups for everyone else on that instance.

The solution: per-tenant serverless instances with a converged engine

Beeda moved the ERP onto MOI Cloud serverless instances — one instance per customer, giving every tenant complete resource isolation. A converged engine carries order creation, inventory lookups and multi-table analysis together.

Storage-compute separation with millisecond autoscaling lets capacity follow the roughly six-hour daily peak, and pay-per-SQL billing replaces always-on licensed capacity. No per-tenant DBA is required.

The outcome: tenants no longer interfere, cost follows usage

Tenant-level resource separation ends the cross-customer interference of hundreds of accounts sharing one SQL Server instance — one customer's report no longer slows another's order entry.

The converged HTAP engine carries order creation, inventory lookups and heavy multi-table analysis together, resolving the read/write contention of the SQL Server setup. For an industry with tight software budgets, per-SQL serverless billing instead of always-on capacity proved decisive.

Solution Architecture

Data sources
  • Dealer orders and inventory
  • Product and dealer master data
  • Reporting and ad-hoc BI queries
  • ~500 tables per customer
MatrixOne Intelligence
  • One serverless instance per tenant
  • Converged HTAP: orders and analytics
  • Storage-compute separation, millisecond autoscaling
  • Pay-per-SQL, no reserved capacity
  • Fully managed, no per-tenant DBA
Business applications
  • Orders without query interference
  • Real-time inventory visibility
  • Complex multi-table reporting
  • Per-customer performance isolation

Related solutions and products