Orleans Design Patterns

• Observer Allows an observer to be notified of any state changes that a grain can choose to publish.
• Cadence Decouple the rhythm of grain interactions from external input using timers.
• Reduce Provides a hierarchical structure to aggregate a value stored in many grains, which would unachievable with a fan-out.
• Smart Cache A performance optimization which uses Orleans as a distributed caching system in front of a storage system. Allows reads to be served from memory, and writes to be optionally buffered.
• Dispatcher A technique to send a batch of messages into Orleans in a single call, and have them distributed internally to the correct grains.
• Hub Provides clients with a single well-known publishing endpoint by channeling events from multiple event sources, dramatically simplifying registration and management of transient subscriptions.

Approximate Performance Expectations

Using X-Large VMs (8 CPU Cores / 14 GB RAM) on Microsoft Azure, with one silo per VM:

• A grain will handle a maximum of 1,000 requests per second.
• A silo will handle a maximum of 10,000 requests per second.
• A silo will hold 100,000 active grains.