VPP Accelerated High Performance & Scalable L3DSR L4 Load Balancer on Top Clos
Delivering fast packets is getting difficult under rapid increase of service endpoints at huge traffic large-scale data center.Especially, time to on-demand deployment of load balancing function neede …
| Talk Title | VPP Accelerated High Performance & Scalable L3DSR L4 Load Balancer on Top Clos |
| Speakers | Naoyuki Mori (Application Engineer, Intel Corporation), Yusuke Tatsumi (Infrastructure Engineer, Yahoo Japan Corporation) |
| Conference | Open Networking Summit North America |
| Conf Tag | |
| Location | San Jose, CA, USA |
| Date | Apr 2- 5, 2019 |
| URL | Talk Page |
| Slides | Talk Slides |
| Video | |
Delivering fast packets is getting difficult under rapid increase of service endpoints at huge traffic large-scale data center.Especially, time to on-demand deployment of load balancing function needed by user-side is becoming shorter as the endpoint changed: bare-metal, VM, to container.To respond to this challenge and to realize networking agility, it is necessary to build up scalable software L4 load balancer cluster with management system.This session will share FD.io VPP (Vector Packet Processing) accelerated scalable L3DSR L4LB system we developed.It provides not only wire rate fast and scalable forwarding performance but also fast deployment and stable operation of LB cluster with zero downtime on Clos architecture in data center.This session will also share what feature we’ve added into VPP to meet practical LB operation requirements, results and how to brush up the system.
