Content-routing as provided by publish/subscribe systems has evolved as a key paradigm for interactions between loosely coupled application components (content publishers and subscribers). Using content-based forwarding rules (also called content filters) installed on content-based routers (also termed brokers), bandwidth-efficiency is increased by only forwarding content to the subset of subscribers who are actually interested in the published content. Software Defined Networking (SDN) is a method that enables installation of filters directly on the TCAM memory of network routers. Compared to traditional broker networks, SDN can significantly reduce the latency until events can be received by subscribers: i) the matching time is significantly reduced, and ii) the communication path can be adapted to directly reflect the underlying network topology and therefore reduce the number of forwarding hops for packets. Initial studies have shown that content-routing protocols based on spatial indexing are very well suited to realize a mapping of filtering operations to header-based packet matching performed in TCAM memory. Filters are represented by identifiers constituting the matching field of flows on switches. However, the limited number of available bits for content representation and the limited number of flows available on the TCAM memory for pub/sub traffic limits the expressiveness of these filters, resulting in false positives or unnecessary traffic in the system. The objective of this thesis, is to design and implement a hybrid pub/sub middleware that allows for filtering of events both on the application layer as well as on the network layer. However, this leads to a trade-off between expressiveness and line-rate performance. In particular, this thesis investigates mechanisms to reduce false positives in the system while maintaining end-to-end latency guarantees at subscribers.