1 / 8

EE807: Software-defined Networked Computing

EE807: Software-defined Networked Computing. KyoungSoo Park Department of Electrical Engineering KAIST. Class Scheduling Is Completed. Everyone is scheduled to lead a class at least once Deawoo Kim dropped out KyoungSoo fills in his place Still, 6 slots are missing

noble-osborne
Download Presentation

EE807: Software-defined Networked Computing

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. EE807: Software-defined Networked Computing KyoungSoo Park Department of Electrical Engineering KAIST

  2. Class Scheduling Is Completed • Everyone is scheduled to lead a class at least once • Deawoo Kim dropped out • KyoungSoo fills in his place • Still, 6 slots are missing • Please volunteer one more time • 5% extra credit (10%: discussion leading) • You can reuse original talk slides for the second talk • Email me with your selection

  3. Final Project • 2-page proposal summary • In the format of Sigcomm poster • Follow the ACM paper format (10pt) • Title, authors, abstract, intro,…, references • On 10/15, 10-minute proposal pitch per team • 3-5 minutes of Q & A • 2-page proposal due 9:00pm on 10/18 • Meet me before your proposal • http://www.ndsl.kaist.edu/~kyoungsoo/calendar

  4. Good Projects • Good motivation • Why do we do this? • Non-trivial problem • Non-trivial, but simple solution • New problem (or new solution) • Don’t repeat the same solution • Better if the space is more general • More general, more impact

  5. More Practical Points • Pick a topic around SDN/NFV/networked systems • Do a literature review on the topic • Read related works • Peek an existing system • If there’s already some system • Get your hands dirty • Some implementation and evaluation • Aim for a workshop or higher venue • Overall quality

  6. Possible Topics • Data plane innovation • SDN separates the control plane from data plane • And assumes data plane is simple – that’s wrong • How do we manage middleboxes using SDN? • User-level networking stacks • Netmap, DPDK, PF_RING, PacketShader I/O engine, etc. • Comparison of usability, performance, etc. • Why? Suggest a better stack?

  7. Possible Topics • Open vSwitch (ovs) • Performance and flexibility • DPDK+ovs is enough for performance? • Can we support flexible routing routing (service)? • Middlebox operation verification • HSA works on L3 networks, can we extend it to L4/L7 services? • Other approaches?

  8. Possible Topics • Software-defined X • X: video, Wi-Fi, or any applications • Distributed systems control through SDN concept • Software-define storage [SOSP’13] • mTCP extension to support virtualization • mTCP is limited to one application • How can we support multiple applications? • Isolation without losing performance?

More Related