VROOM: Virtual ROuters On the Move

Yi presents:

• How much it would help if protocols are perfect with recovery? Answer: VROOM is generic. VROOM is not designed to address a particular problem, it is anew way of thinking. Even if convergence is fast, in case of taking routers off-line, we have to reconfigure, in this case we can avoid that.
• Can VROOM guarantee that the protocols don’t re-converge? Answer: VROOM is transparent to control plane.
• What is the deployment model? Do we need to change the infrastructure? No, the infrastructure is there.
• Is this indirection level just makes things more complex? No, VROOM provides cleaner interfaces for network management.
• Are we saving the power with this approach? There is certainly an overhead, but VROOM can save power, because it adapts to daily usage changes.

General discussion:

• The power might be saved if we approach it from sociological point of view.
• Standby might not be a solution as it consumes the energy anyway...
• VROOM: Why don’t we use FRR for maintained downtown instead of VROOM? Answer: If we look at the core routers, we can survive using FRR, however if it is the edge router then we offer the advantage. We also offer other applications.
• VROOM: Can we just get rid of routing protocols? Answer: We still need benefits of statistical multiplexing...

• Question to Mark: Is agent the right solution? Answer: yes, the industry is moving towards it...

• Alison: I’m intrigued that we have three different approaches. What is the impact to the energy savings if we put all these approaches together in practice? All approaches + user awareness would save a lot.