Study in Hybrid Networks

1. Telekinesis: Controlling Legacy Switch Routing with OpenFlow in Hybrid Networks. In SIGCOMM SOSR (2015).

　　1.1 keywords: link update, L2, MAC learning, fake packets

　　1.2 reference: 

　　　　1.2.1 Reaping the Benefits of Incremental SDN Deployment in Enterprise Networks. In USENIX Annual Technical Conference (2014).

　　　　　　　　　an optimization framework to determine the location of partial SDN deployment and enforcing every network flow traverse at least one SDN-enabled switch

　　　　1.2.2 HybNET: Network Manager for a Hybrid Network Infrastructure. In Middleware (2013).

　　　　　　　　　providing a common configuration mechanism for hybrid networks and automatically translates legacy to SDN configurations and vice versa. HybNET proposes a netconf-based mechanism to change legacy switch routing but its speed (more than 10s to configure a switch) makes it incompatible with real world networks.

　　1.3 questions:

　　　　1.3.1 how to avoid an error network status caused by the fake packet? 

　　1.3 my future work:

　　　　1.3.1 thinking out a scene of link update, such as link/node failure 

　　　　1.3.2 consistent update

2. Congestion-Aware Single Link Failure Recovery in Hybrid SDN Networks. In INFOCOM (2015).

　　1.1 keywords: pre-configured IP tunnels, L3, candidate table construction and greedy approach, LB

　　1.2 reference:

　　　　1.2.1 Opportunities and research challenges of hybrid software defined networks. In ACM SIGCOMM Computer Communication Review (2014).

　　　　1.2.2 Traffic engineering in software defined networks. In INFOCOM (2013).

　　　　　　　　  dynamically adjust routing decision based on measured traffic in hybrid networks with a few sdn switches.

　　　　1.2.3 Safe updates of hybrid sdn networks. In UCL, Tech. Rep (2013).

　　　　　　　　  https://www.baidu.com/link?url=fZWgF3O_MhD0akWWUWSU-LXIQnFN5BeN6EdkxzJAwqes5lFpwaQLEUKoU-N7UrPKamZsyWBMR8w4I_DURUWMeK&wd=&eqid=adb337b00003125e000000035670232e

　　　　　　　　  designing a safe and consistent rule update mechanism in a hybrid SDN network to avoid violating security policies.

　　1.3 questions:

　　　　1.3.1 how does the controller get the status of each link?

　　　　　　　　  They make the same assumption as [1.2.2] that each router runs protocols such as OSPF-TE which allows the SDN controller to gather link load information in the network. In OSPF-TE, information such as available bandwidth can be exchanged using opaque LSA, so each SDN switch is able to get the link utilization of all the links in the network and forward this information to the SDN controller.

　　1.4 my future work:

　　　　1.4.1 transfering the scene of article 2 to L2 networks, and making a consistent update

　　　　1.4.2 My thought about consistent updating should be based on a newest artical.