Programmable Congestion Control
Peyman Teymoori, University of Oslo
SDN & OpenFlow World Congress,
The Hague, October 2016
Congestion Control in the Internet
• Problems with the Internet:
– TCP scalability with:
• The diameter of the network
• The bottleneck link capacity
• The number of flows
– Different link types
• One end-to-end controller!
– Split-TCP (PEPs):
• IPsec and SSL
• Scalability with the number of flows
• Processing delay at splitters
– PEP-less solutions:
• Again hacking!
Not by hacking/patching
• RINA’s benefits for congestion control:
– breaking up the long control loop into shorter ones,
– controlling flow aggregates ,
– deployment of arbitrary congestion controllers per DIF.
• The benefits occur
– just as the result of layering,
– without changing a line of code in the mechanism itself,
– without the side effects of Internet improvements (e.g. Split-TCP)
– just by different DIF configurations
• It’s a matter of configuration/policy! So, programmable…
– Horizontal: Consecutive DIFs
– Vertical: Stacked DIFs
– Around: In-Network Resource
– Use of customized controllers per DIF
Current Status & Further Work
• Developing a new congestion controller policy, called LGC:
– based on the Logistic Growth function,
– taking advantage of RINA’s flow aggregation,
• The controller can be smoothed.
– playing with it in different use cases, e.g. data center.
• Evaluating multiple control loops.
Further information can be found here.