A Scalable Protocol for Content-Based Routing in Overlay Networks
NCA '03 Proceedings of the Second IEEE International Symposium on Network Computing and Applications
A data-oriented (and beyond) network architecture
Proceedings of the 2007 conference on Applications, technologies, architectures, and protocols for computer communications
Adaptive content-based routing in general overlay topologies
Proceedings of the 9th ACM/IFIP/USENIX International Conference on Middleware
Field-split parallel architecture for high performance multi-match packet classification using FPGAs
Proceedings of the twenty-first annual symposium on Parallelism in algorithms and architectures
Proceedings of the 5th international conference on Emerging networking experiments and technologies
Architectures for the future networks and the next generation Internet: A survey
Computer Communications
On content-centric router design and implications
Proceedings of the Re-Architecting the Internet Workshop
Hi-index | 0.00 |
The future Internet is expected to be a highly intelligent that can route packets not only with an explicit destination address but also with the content. This paradigm shift in the network architecture from host- to content-centric communication naturally leads to the contemplation of the shift in the network layer devices, i.e. routers. In other words, the hardware architecture of routers should also be able to support content-centric communication. In this paper, we propose a new router architecture to manage a large information of contents and large-scale number of users. In order to complete the packet forwarding within the network layer, routers acting as the brokers of a publish/subscribe system should maintain the information of content names and the subscribers. We propose three memory structures for name lookup tables in routers, each with a different combination of memory types depending on the usage purpose. The proposed memory architecture is evaluated with parameters such as memory cost, latency, and utilization using real-life and synthetic databases that have a Zipf distribution. We show the memory architecture which has the lowest manufacturing cost and the lowest latency for storing the given database within a fixed budget.