Fewer, larger conversations
The services are designed to answer useful questions in fewer requests, which reduces chatter between the storefront and the engine.
A compact services layer, separable application tiers and an approach designed for catalogs that keep growing.
The services are designed to answer useful questions in fewer requests, which reduces chatter between the storefront and the engine.
Services, administration and the scheduler can be distributed across servers when load requires it.
Performance work is about measuring the real catalog and order mix, not only the size of the product table.
Slow storefronts are often caused by chatty templates, unindexed queries or oversized media, not by the commerce engine alone. Measure the page, the service call and the database separately before changing the architecture.
Caching helps when the same data is requested repeatedly. It hurts when the cache is filled with low value variants. Start with a small, explainable cache policy.
Use a catalog size and order pattern that resembles the business. A benchmark on a tiny sample can flatter every layer of the stack. The useful question is how the system behaves at the volumes the team expects next year.
Standards, documentation and tools that are relevant to this topic. They open on the publisher's site.
The useful test is a real assortment, a real market and a real order. Bring those to a conversation rather than a generic feature list.