This section discusses the distributed aspects of ONOS's network topology representation.

Overview 

ONOS, as As part of its north bound northbound API, ONOS provides applications with access to a global network topology view. Applications operate on this view to perform various network functions such as path computation , flow provisioning, etc.

Background

and flow provisioning, among others.ONOS is built from the ground up as a distributed SDN operating system. Consequently, ONOS is deployed as a collection of servers that coordinate with each other to provide capabilities that is greater than the sum of its parts. For example, distribution provides fault-tolerance and resilience even when individual controller instances fail. Additionally, the system as whole can take on work-load that is far greater than what a single instance might be able to handle (scalability).   

Global Network Topology View

While distribution has its benefits, it also presents some interesting engineering problems. Foremost among them is maintaing maintaining the consistency of state deployed across various ONOS instances. This is especially evident in the case of global network topology state management where each controller has to expose a view of the entire network even though at any given point in time it only has direct visibility over a subset of the network.  
There are some properties that a good network topology state management solution ought to have in a distributed setting:
 

1. Completeness: Even though each controller only has direct visibility and influence over a subset of the network, they should all work together to ensure each controller’s network topology view reflects the state of the entire network.
2. Accuracy: Switches, ports and links go up and down. Each controller’s network view should always expose the correct state for various network elements. This also means that each controller’s network view should quickly change to reflect any changes in the underlying network.
3. Low latency access: Network topology view is a heavily consumed piece of state and therefore it is very important that the chosen mechanism provide low latency access the network view.

...

Approach

Terminology  

In ONOS, the entire global network topology state is cached in memory on each instance. This provides applications low latency access to topology state. Before we go into the details of how the topology state is kept sync across instances and with the state of the physical network, it is useful to define couple of concepts: 

...