Scalability of Software-Defined Networking
This study discusses the scalability challenges in Software-Defined Networking (SDN) and proposes solutions to improve controller scalability, reduce flow initiation overhead, and enhance resilience to failures. Various techniques, such as distributed systems and parallel processing, are explored to address these scalability issues.
Uploaded on Apr 04, 2025 | 0 Views
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
On Scalability of Software- Defined Networking S.H. Yeganeh, A. Tootoonchian, Y. Ganjali University of Tronto IEEE Communication Magazine 1
Motivations There have always been concerns about performance and scalability about SDN Argue that there is no inherent bottleneck to SDN scalability Often came from implicit and extrinsic assumption These problems are also faced by traditional networks If carefully implemented, these problems can be avoided, ex: implemented as distributed system 2
Discussed Scalability Issue Scalability of controller (major) Flow Setup overhead Resilience to failures 3
Controller Scalability Early benchmark of NOX Only handle 30K requests while maintaining a reasonable flow install time (sub-10ms) Two way to solve this problem Improve the parallelism, improve IO performance Implement as distributed system 4
Controller Scalability Improve the parallelism, improve IO performance Parallel version of NOX (HOT-ICE 2012) DIFANE (SIGCOMM 2010) DevoFlow (SIGCOMM 2011) Implement as distributed system Onix (OSDI 2010) HyperFlow (INM/WREN 2010) Kandoo (HotSDN 2012) 5
Flow Initiation Overhead Early design such as Ethane and NOX makes people think flows have to be reactive This is not true Two possible way Aggregate rules Proactively install rules DIFANE in some way did this 6
Resiliency To Failures Early design use only one centralized controller This doesn t have to be the case Link failure notification Traditional: flooded across the network SDN: switch to controller (info propagation delay is no worse, and a controller dedicated to deal with the problem) Not worse than traditional network Similar techniques can be applied, ex: preprogrammed backup paths 8
Scalability In Different Settings Data Centers Large number of devices, controller will be overload Use proactive install rules but sacrifice precision and reactivity Use distributed system if those can t be sacrifice Service Provider Networks Geographically distributed, worsen the latency issue Partition into regions High number of flows Aggregated rules 9
Other Topic Behavioral and Programming Abstractions Ease of management Testing and Verification Ease of troubleshooting Extensibility How to support more protocols 10
Conclusion Argue the scalability concern Are neither caused by nor fundamentally unique to SDN Can be addressed without losing the benefit of SDN Emphasize the potential gain on other topics 11
