TY - GEN
T1 - Controlling software router resource sharing by fair packet dropping
AU - Addanki, Vamsi
AU - Linguaglossa, Leonardo
AU - Roberts, Jim
AU - Rossi, Dario
N1 - Publisher Copyright:
© 2018 IFIP
PY - 2018/1/1
Y1 - 2018/1/1
N2 - We demonstrate a practical way to achieve multiresource sharing in a software router, where both bandwidth and CPU resources may be bottlenecks. Our main idea (published in a same-titled paper in this year IFIP Networking conference [1]), is to realize per-flow max-min fair sharing of these resources by wisely taking drop decisions according to the state of a shadow system. We implement our FairDrop proposed algorithm in Vector Packet Processor (VPP), a novel high-speed software router architecture. We demonstrate FairDrop is capable of fairly sharing CPU cycles among flows with heterogeneous computing workload, at 10Gbps on a single core.
AB - We demonstrate a practical way to achieve multiresource sharing in a software router, where both bandwidth and CPU resources may be bottlenecks. Our main idea (published in a same-titled paper in this year IFIP Networking conference [1]), is to realize per-flow max-min fair sharing of these resources by wisely taking drop decisions according to the state of a shadow system. We implement our FairDrop proposed algorithm in Vector Packet Processor (VPP), a novel high-speed software router architecture. We demonstrate FairDrop is capable of fairly sharing CPU cycles among flows with heterogeneous computing workload, at 10Gbps on a single core.
M3 - Conference contribution
AN - SCOPUS:85127951476
T3 - 17th International IFIP TC6 Networking Conference, Networking 2018
SP - A9-A10
BT - 17th International IFIP TC6 Networking Conference, Networking 2018
PB - IFIP
T2 - 17th International IFIP TC6 Networking Conference, Networking 2018
Y2 - 14 May 2018 through 16 May 2018
ER -