Using the filtering-based routing algorithm as a starting point, develop an alternative algorithm that illustrates how the use of advertisements can result in significant optimization in terms of message traffic generated.Using the filtering-based routing algorithm as a starting point, develop an alternative algorithm that illustrates how the use of advertisements can result in significant optimization in terms of message traffic generated.

What will be an ideal response?

The filtering-based routing algorithm significantly reduces the amount of traffic generated in comparison to flooding approaches by maintaining state information on each node in the overlay. In particular, each node must maintain a list of all directly connected subscriptions and also routing information that determines if there is a valid (matching) subscription down a connected path. This means that publish events are only forwarded down paths that will lead to a valid subscription, thus reducing the amount of traffic generated specifically by publish events. The downside is that subscription events need to be propagated fully through the overlay for this to be optimal and this flooding leads to a significant amount of traffic, an overhead which can be reduced by advertisements.
An advertisement-based approach applies the same principle to subscriptions as the above scheme applies to publish events. Advertisements are issued by publishers to indicate an intent to publish particular kinds of events in the future. Each node must now hold an additional advertisement-based routing table. In the same way that the initial (subscription-based) routing table optimizes the forwarding of publish events (by ensuring there is a valid path to a matching subscription), the second advertisement-based routing table is used to optimize the routing of subscriptions. In particular, subscriptions only need to be forwarded only into parts of the network where publishers have issued overlapping advertisements. This approach is described in detail in [Muhl et al. 2006].