Centralized Routing vs Distributed Routing | Centralised Routing vs Distributed Routing
Routing is the process of choosing which paths to be used to send network traffic, and sending the packets along the selected sub-network. Centralized routing model is a routing model in which routing is centrally carried out using a centralized database. On the contrary, distributed routing model is a routing model, which deals with carrying out routing using a distributed database.
What is Centralized Routing?
Centralized routing model is a routing model in which routing is centrally carried out using a centralized database. In other words, the routing table is kept at a single “central” node, which should be consulted when other nodes need to make a routing decision. This centralized database possesses a global network view. Centralized routing is most appropriate in specific domains with systems providing DWDM (Dense Wavelength Division Multiplexing) transmission. The reason for this is that these DWDM systems contain OADM (Optical Add-Drop Multiplexer) that can be reconfigured, within the starting point and end point of the communication medium. Proponents of centralized routing suggest that because most information such as details on SRLG (Shared Risk Link Group) and parameters of performance do no change frequently (and these information may never be self-discovered or advertized), they are ideal to be resided in a central database. In a centralized model, information of state can easily be accessed. Therefore, the information on dependencies (with regard to routing) between circuits (to make sure the diversity exists) can be handled relatively easily when terminals are not shared between circuits, and this is ideal for a centralized routing model. The centralized model uses global state information. The computations that are carried out (such as pre-computing the paths for restoring) can be highly benefitted from this global information and therefore this is suitable for a centralized model.
What is Distributed Routing?
In distributed routing model, each node keeps a separate routing table. Distributed routing model is a routing model, which is excellent for domains that can be identified as entirely opaque. The main reason for this is that constraints on impairments do not play any role in routing within these aforementioned domains. In case of a failure (when there is a need to restore rapidly), distributed routing system can be relied upon to bear the responsibility of the on-demand computation of a paths of recovery for each of the light-paths that have failed (even at the point of detecting an expected failure). Finally, the distributed routing model is highly consistent with the existing Internet’s own distributed routing philosophy.
What is the difference between Centralized Routing and Distributed Routing?
Centralized routing model carries out routing using a centralized database, while distributed routing model deals with carrying out routing using a distributed database. In simple terms, one central node holds the routing table in the centralized model, while each node keeps a routing table in the distributed model. Because much of the information does not change frequently, many believe that this information is much suitable to be resided in a centralized database. The pre-computations needed for restoration can take the advantage of global information available in a centralized database. But, unlike the distributed routing system, the centralized system cannot be relied upon to bear the responsibility of the on-demand computation of paths of recovery for each of the light-paths that have failed (at a time of detecting an expected failure). Unlike the centralized approach, the distributed routing model is highly consistent with the existing Internet’s own distributed routing philosophy.