Michael Cahill, Director of Engineering (Storage), MongoDB
When the WiredTiger storage engine was created, the use case we had in mind was applications with modest numbers of collections. That led to various choices during the design, such as storing each collection in a separate file. However, MongoDB customers have an enormous variety of use cases, including multi-tenant applications where each user has a separate database and each database contains hundreds of collections.
To support these applications efficiently, we have evolved the storage layer with better testing, better analysis tools, and more scalable data structures and algorithms. This session will explain how we can now run workloads with over a million collections.