How long do your SDN controllers take to converge?
A key latency metric for network control plane systems is convergence time: the duration between when a change occurs in a network and when the network has converged to an updated configuration that accomodates that change. The faster the convergence time, the better.
Convergence time depends on many variables: latencies between network devices, the number of network devices, the complexity of the replication mechanism used (if any) between controllers, storage latencies, etc. With so many variables it can be difficult to build an intuition for how the variables interact to determine overall convergence time.
The purpose of this tool is to help build that intuition. Based on analytical models of communication complexity for various replication and network update schemes, the tool quantifies convergence times for a given topology and workload. With it, you can answer questions such as "How far will my current approach scale while staying within my SLA?", and "What is the convergence time of my network under a worst-case workload?".
Accuracy - Monte Carlo simulation means we run repeated trials for a given configuration. This parameter determines how many samples we take to determine the CDF.
Operation Latency - How long does each processing step take? Here, this includes the time to compute a write response at a single node (W), the network latency between controllers (Pc) and between switches and controllers (Ps), and the latency to compute a read response at a single node (R). The distributions are currently delayed according to a Pareto distribution. Lower alpha correspond to heavier tails.
Topology Generation - Controls the arrangement and number of controllers and switches in the network.
|Traditional||0 ms||0 ms|
|Single||0 ms||0 ms|
|1PC||0 ms||0 ms|
|2PC||0 ms||0 ms|
|Paxos||0 ms||0 ms|