requirements:

• Generality
• Scalability
• Reliability
• Simplicity
• Control plane performantce

contributions:

• First, Onix exposes a far more general API than previous systems
• Second, Onix provides flexible distribution primitives (such as DHT storage and group membership) allowing application designers to implement control applications without re-inventing distribution mechanisms

• Physical infrastructure
• Connectivity infrastructure
• Onix
• Control logic

• Onix’s API consists of a data model that represents the network infrastructure, with each network element corresponding to one or more data objects.
• The copy of the network state tracked by Onix is stored in a data structure we call the Network Information Base (NIB)
• While Onix handles the replication and distribution of NIB data, it relies on application-specific logic to both detect and provide conflict resolution of network state as it is exchanged between Onix instances, as well as between an Onix instance and a network element.

Onix supports three strategies that can used to improve scaling.

• Partitioning: a particular controller instance keeps only a subset of the NIB in memory and up-to-date
• Aggregation: one instance of Onix can expose a subset of the elements in its NIB as an aggregate element to another Onix instance.
• Consistency and durability: The control logic dictates the consistency requirements for the network state it manages.

Control applications on Onix must handle four types of network failures:

• forwarding element failures,
• link failures,
• Onix instance failures,
• failures in connectivity between network elements and Onix instances (and between the Onix instances themselves

For network state needing high update rates and availability,
Onix provides a one-hop, eventually-consistent,
memory-only DHT