White Papers

Understanding OPNFV

Issue link: https://read.uberflip.com/i/1344850

Contents of this Issue


Page 80 of 144

Understanding OPNFV 80 OPNFV Project Analytics Success Story: OPNFV Doctor with OpenStack Vitrage An example of a successful OPNFV project creating carrier-grade features in upstream projects is Doctor. The project was proposed by NTT DOCOMO in November 2014 to ensure that fault management in the virtualized world would be just as good as in the physical world. At the time of the project proposal, fault management in physical networks was very well understood and efficiently implemented. A failure would get detected by the monitoring system, and automatic failover would take place. However, the combination of virtualization and cloud-hosted VNFs broke this mechanism. Given the state of OpenStack at the time, virtual machine, storage or network port failures would eventually get detected using software polling or timeouts, but while these long durations were acceptable in enterprise use cases, they were unacceptable in NFV. The Doctor project started to address this by working with numerous upstream OpenStack projects such as Nova, Neutron, Congress and Aodh. The goal of OPNFV Doctor was straightforward: once an infrastructure failure is detected, map it to the affected virtual resources, and take failover or any other action within 1 second. The challenges were numerous: ● Defining what infrastructure failure is ● Mapping it to affected virtual resources

Articles in this issue

Links on this page

view archives of White Papers - Understanding OPNFV