Deep Learning Forwarding in NDN With a Case Study of Ethernet LAN

In this paper, the authors propose a novel forwarding strategy based on deep learning that can adaptively route interests/data packets through ethernet links without relying on the FIB table. The experiment was conducted as a proof of concept. They developed an approach and an algorithm that leverage existing intelligent forwarding approaches in order to build an NDN forwarder that can reduce forwarding cost in terms of prefix name lookup, and memory requirement in FIB simulation results showed that the approach is promising in terms of cross-validation score and prediction in ethernet LAN scenario.

Route Prefix Caching Using Bloom Filters in Named Data Networking

This paper proposes an elaborate route prefix caching scheme for fast packet forwarding in named data networking (NDN) which is a next-generation Internet structure. The name lookup is a crucial function of the NDN router, which delivers a packet based on its name rather than IP address. It carries out a complex process to find the longest matching prefix for the content name. Even the size of a name prefix is variable and unbounded; thus, the name lookup is to be more complicated and time-consuming. The name lookup can be sped up by using route prefix caching, but it may cause a problem when non-leaf prefixes are cached. The proposed prefix caching scheme can cache non-leaf prefixes, as well as leaf prefixes, without incurring any problem. For this purpose, a Bloom filter is kept for each prefix. The Bloom filter, which is widely used for checking membership, is utilized to indicate the branch information of a non-leaf prefix. The experimental result shows that the proposed caching scheme achieves a much higher hit ratio than other caching schemes. Furthermore, how much the parameters of the Bloom filter affect the cache miss count is quantitatively evaluated. The best performance can be achieved with merely 8-bit Bloom filters and two hash functions.

Scalable Name Lookup for NDN Using Hierarchical Hashing and Patricia Trie

Named data networking (NDN) is a content-centric network for the future of the internet. An NDN a packet is delivered based on a content name instead of a destination IP address. The name lookup for packet forwarding is challenging because the content name is variable, and its space is unbounded. This paper proposes a novel name lookup scheme that employs a hashing technique combined with Patricia tries. In this scheme, hash tables are dynamically maintained according to the hierarchical structure of the name, so the name can effectively accommodate variable and unbounded content names. Unlike chaining, a Patricia trie compares a key string only once at a leaf node, so it provides a fast name lookup. The proposed lookup scheme is implemented and evaluated by using an Intel Core i7-2600 CPU and 4GB of memory. Experimental results show that our scheme gives good scalability and a high throughput of name lookup with a controlled memory consumption.

Towards Pending Interest Table Management Solutions in Named Data Networking

The transition of current Internet from host-centric communication to content-centric communication has been realized due to rapid growth in the number of users and their demands. Named Data Network (NDN) has emerged as Future Internet Architecture to meet the demands related to content oriented communication. Among other data structures such as Forwarding Information Base (FIB) and Content Store (CS), Pending Interest Table (PIT) holds record of the pending interests for which response has not yet received. Despite the benefits of PIT for fast packet forwarding (such as address less communication, loop detection, multi-path routing, packet aggregation, efficient bandwidth utilization), there are challenges faced by PIT. These challenges include name lookup speed, PIT memory consumption and PIT Entry Lifetime (PEL) that may render PIT as a bottleneck and hence will affect the system performance. In this paper, we thoroughly and systematically study the PIT management schemes. Our study covers the solutions for PIT management spanning over the last 7 years. As far as considering our learning, we are the first to study PIT based on both PIT memory size reduction to achieve fast look-up speed and different PEL calculation schemes. The study will help the research community in modeling better and optimized PIT management solutions. The paper also presents the research trends on the basis of different regions, publishing outlets, contributing patterns etc. in the field of PIT management in NDN.