White Papers

Understanding OPNFV

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

Contents of this Issue


Page 40 of 144

Understanding OPNFV 40 multiple teams. Organizational structure readiness questionnaire: Question Readiness: 1-5 (1 = Not ready, 5 = Completely ready) Has there been any example where cross-functional groups formed teams for new technology? Can you imagine completely autonomous teams per NFV service or component (such as NFVI, VIM, MANO) that owns dev, test, release and production? Can you imagine a decentralized decision-making approach where each team, within reason, could make their own technical decisions? Can you imagine the communication between NFV services teams and the platform team to be completely automated? Are there mechanisms to discuss requirements that affect multiple teams, and to coordinate these requirements? Impact to Process During the 2016 OpenStack Summit in Austin, Jonathan Bryce talked about a customer who took 44 days to deploy an application. Once the company implemented OpenStack, they were able to cut this down to a disappointing 42 days. They had left all the checkpoints and manual signoffs intact! Once the cultural issues and processes were addressed, the duration was cut to 2 hours. This proves that to succeed, organizations need "software defined people". Processes may range from the mundane, such as checkpoints and signoffs, to more complex, such as inventory management, security, service assurance, auditing, and so on, which are all difficult in a virtual world as instances spin-up and down. Often there is a lot of bureaucracy in place for any change, put in place to promote stability and reduce risk. With NFV, a complete review of all processes will be required, and those that do not truly contribute to stability will have to be replaced based on a culture of trust and accountability. Any spreadsheet, meeting,

Articles in this issue

view archives of White Papers - Understanding OPNFV