A novel multi-path traffic control mechanism in named data networking

Fuzzy Logic‐based Efficient Interest Forwarding (FLEIF) in Named Data Networking

“…All the packets are transferred with great efficiency as duplicated packets may be discarded at the base node to prevent congestion . For multipath forwarding, Li et al adopted convex optimization between different faces in order to find the appropriate rate limit for Interest forwarding in order to avoid congestion. Proactive congestion control can also overcome congestion problem as described in the work of Zhou et al Explicit Congestion Notification is combined with sending rate of Interest to attain the optimization for the traffic forwarding and to avail the best rate for individual flow.…”

“…A number of congestion control mechanisms [6][7][8][9][10][11][12][13][14][15][16][17] have been proposed for CCN/NDN and many of them are under research 1,[18][19][20][21][22][23][24][25][26][27][28][29][30][31][32] to improve the performance of the network by making the system more reliable and faster. Some studies 5 suggest to request a large block of data in a single Interest packet, in order to avoid traffic caused by multiple Interests.…”

A Practical Congestion Control Scheme for Named Data Networking

“…Moreover, PCON's forwarding strategy converges to the optimal split ratio, even in the case where congestion signals result Our implementation of HBH Interest shaping is arguably simplistic, as it doesn't even try to predict the real bottleneck bandwidth and instead uses the incorrect value of the local link capacity. However, the related work we know of ( [5,22,19,10,27,26,17,11]) doesn't consider the possibility of unknown link bandwidth, and leaves the behavior in this case unspecified. Moreover, the task of explicitly estimating the bandwidth of an IP tunnel or wireless link is non-trivial.…”

“…Another approach to NDN congestion control is Hop-by-Hop (HBH) Interest Shaping [5,22,19,10,27,26,17,11], based on the assumption that a router can control the amount of returning data on its links by controlling the Interest sending rate in the opposite direction. As shown in Section 4.6, these schemes require explicit knowledge of the available link bandwidth, which limits their effectiveness for wireless networks and IP tunnels.…”

“…NDN's stateful forwarding plane enables routers to control congestion at each hop, by either dropping Interest packets or diverting them to alternative paths. However, most existing solutions in this direction (HBH Interest Shaping) [5,22,19,10,27,26,17,11] assume known or predictable link bandwidths and Data chunk sizes, which limits their effectiveness in scenarios where these assumptions do not hold true. For example, the largest experimental NDN deployment, the NDN Testbed [2], runs over UDP tunnels where the underlying bandwidth is unknown and keeps changing; a video-ondemand provider may respond to congestion by adjusting the video quality (hence Data chunk size) dynamically.…”