Title

Make Item Ledger Entry and Value Entry location-aware to improve scalability for multi-store and warehouse environments

Description

Item Ledger Entry (ILE) and Value Entry (VE) are primarily designed to manage inventory by location (store / warehouse).

However, their current physical design is optimized for single-location companies and does not scale efficiently in multi-location, high-concurrency environments.

In companies operating many stores and warehouses (for example, 50 retail stores and multiple distribution centers), most inventory transactions are:

• strictly scoped to a single location,
• independent from other locations,
• yet forced to share the same global physical structures (indexes, aggregations, write paths).

This results in:

• unnecessary contention,
• lock waits,
• deadlocks between transactions that are not functionally related from a business perspective.

Same root issue, two common scenarios

Retail multi-store

• High-volume sales and replenishment postings per store
• Availability calculations almost always filtered by Location Code
• Stores are operationally independent but technically competing for the same resources

Warehouses and distribution centers

• High-throughput inbound, outbound, and transfer postings
• Concurrent processes such as picking, shipping, and replenishment
• Contention with store activity despite clear business separation

Although the processes differ, the root cause is the same:

independent inventory domains are forced to share a single global physical structure.

Proposal

Introduce an optional, Microsoft-managed performance profile that makes Item Ledger Entry and Value Entry location-aware.

In this mode:

• inventory reads and writes are physically isolated by location,
• transactions from one location no longer block those from other locations,
• the physical data layout reflects the actual business scope of inventory.

Implementation details (physical layout, partitioning, indexing strategy) should remain internal to the product, with the sole objective of reducing cross-location contention while preserving the existing functional model.

Why this change is needed

• ILE and VE are core system tables; partners cannot safely adapt their physical design
• Current workarounds (serialization, job queues, reduced parallelism) lower overall throughput
• Multi-store retail and warehouse automation scenarios are becoming standard, not edge cases

Acceptance criteria

• Measurable reduction in lock waits and deadlocks on ILE/VE
• Improved posting throughput when multiple locations operate in parallel
• No impact for single-location companies
• Opt-in, reversible, and observable through telemetry

Conclusion

The current design of ILE and VE is suitable for simple environments but limits scalability in multi-location businesses.

Making these tables location-aware, through an optional performance profile, would align the technical architecture with the business reality of inventory management.