WHL—reading notes《A Native Content Discovery Mechanism for the Information-Centric Networks》

2017-11-11

Posted by 林伟宏

Main goal: discovering content in the caching nodes of an AS to reduce costly inter-AS traffics.

 

Two broad categories of exist approaches:

  • searching content opportunistically (on-path) along the traditional shortest path toward content origin, such as [11,30], which gains is limit;
  • coordinated content discovery approaches: requests are forwarded to some designated caching nodes that are likely to hold this content, such as[3,20,22], which have higher gains at price of higher overhead to maintain up-to-date state of the cached content.

 

Main work in this paper: try to combine the merits of tow above broad categories of exist approaches by using an opportunistic coordination approach, which requires each router to keep track of only a small amount of state and does not require signaling or update protocol to locate content.

 

To that end, enhance the NDN routing architecture with a new component called Ephemeral Forwarding Information Base (EFIB), which keeps track of the direction in which data chunks were temporarily cached in the recent past.

 

An EFIB entry is triggered by a returning Data packet and points towards the direction (i.e., next hop) in which data chunks were temporarily cached in the recent past.

Challenge: The introduction of the EFIB table requires new forwarding strategies for next-hop selection for the matching EFIB and FIB entries, with trade-off s in retrieval latency and overhead in terms of duplicate Data packet responses and circulated Interest packets.

To solve this challenge, they propose a budget-based multicast forwarding strategy, where each Interest packet is given a “forwarding budget” that can be spent (at the intermediate routers) on forwarding the Interest on-path towards the content origin and/or sending o -path interests towards caches.