Middlebox Discovery for TCP Networks

Middlebox Discovery for TCP Networks
paly

This article discusses the draft knutsen tcpm middlebox discovery 03, which defines a new TCP option for in-band discovery of middleboxes in TCP networks. It covers the benefits of this option, how it is designed for easy adoption, and includes practical implementation experience and suggestions from the tcpm mailing list.

  • Uploaded on | 0 Views
  • jameson jameson

About Middlebox Discovery for TCP Networks

PowerPoint presentation about 'Middlebox Discovery for TCP Networks'. This presentation describes the topic on This article discusses the draft knutsen tcpm middlebox discovery 03, which defines a new TCP option for in-band discovery of middleboxes in TCP networks. It covers the benefits of this option, how it is designed for easy adoption, and includes practical implementation experience and suggestions from the tcpm mailing list.. The key topics included in this slideshow are Middlebox Discovery, TCP Networks, TCP Option, TCPM, Implementation Experience,. Download this presentation absolutely free.

Presentation Transcript


1. Middlebox Discovery Jamshid Mahdavi Andrew Knutsen March 23, 2010

2. Talk Outline Middlebox Discovery ID Summary and Status Discussion of Middlebox Needs Other Common Middlebox Issues of Potential Interest to IETF References

3. ID Summary draft-knutsen-tcpm-middlebox-discovery-03 Defines a new TCP Option for in-band discovery of middleboxes Designed from the ground up to: Consume only a single TCP Option Kind for all vendors who need this capability Allow for safe proprietary use as well as future standardized use Includes lessons from years of practical implementation experience Incorporates numerous good suggestions from tcpm mailing list

4. ID Status Working Group has chosen not to take this up as a WG item Draft has been submitted for IESG approval

5. Evolving Internet Connectivity 1980s: Direct IP to IP connections 1990s: Firewalls and NATs become prevalent on nearly all paths 2000s: Increasing use of higher level middleboxes Proxies (caching, security) Access points Acceleration devices Load balancers Rate shaping / TCP enhancing devices

6. What about End-to-End Arguments? David D. Clark, Marjory S. Blumenthal, Rethinking the design of the Internet: The end to end arguments vs. the brave new world, August 10, 2000. Paper outlines many requirements that we see today

7. Todays Drivers Security Cybercrime and malware are growing problems Performance Bandwidth savings via advanced compression technologies Latency savings via protocol optimizations Improved goodput via TCP optimizations New emerging market for proxies as IPv6 transition appliances

8. Known Problems There are a few problems we see all the time which the IETF could have an impact on: TCP ACK storms Application Networking devices often use fail-to-wire bridging If fully transparent, when failure happens, ACK storm ensues Asymmetric routing (or routing changes) Often cited as a key reason transparent intercept is incompatible with Internet architecture But vendors have numerous proprietary solutions to handle this Amplification of known issues PMTU black holes Broken support for RFC1323 and other extensions to TCP and IP

9. References (1/3) Historical references on proxies and Internet architecture: Chatel, M., Classical versus Transparent IP Proxies, RFC1919 (1996). http://datatracker.ietf.org/doc/rfc1919/ Saltzer, J. H.; Reed, D. P.; Clark, D. D., End-to-End Arguments in System Design. (1984). http://web.mit.edu/Saltzer/www/publications/endtoend/endtoend.pdf Clark, D. D.; Blumenthal, M. S., Rethinking the design of the Internet: The end to end arguments vs. the brave new world. (2000). http://cyberlaw.stanford.edu/e2e/papers/TPRC-Clark-Blumenthal.pdf Clark, D. D.; Sollins, K.; Wroclawski, J.; Faber, T., Addressing Reality: An Architectural Response to Real-World Demands on the Evolving Internet. (2003). http://www.isi.edu/newarch/DOCUMENTS/Principles.FDNA03.pdf

10. References (2/3) Research publications: Spring, N. T.; Wetherall, D., A Protocol Independent Technique for Eliminating Redundant Network Traffic. (2000). http://www.cs.umd.edu/~nspring/papers/sigcomm2000.ps.gz Li, Q., A Novel Approach to Manage Asymmetric Traffic Flows for Secure Network Proxies. (2008). http://www.springerlink.com/content/13n10l6u011530t1/ Anand, A.; Gupta, A.; Akella, A.; Seshan, S.; Shenker, S., Packet Caches on Routers: The Implications of Universal Redundant Traffic Elimination. (2008). http://ccr.sigcomm.org/online/files/p219-anand.pdf Anand, A.; Sekar, V.; Akella, A., SmartRE: An Architecture for Coordinated Network-wide Redundancy Elimination. (2009). http://ccr.sigcomm.org/online/files/p87.pdf 10

11. References (3/3) Vendor references: Salchow, K. J., Load Balancing 101: The Evolution to Application Delivery Controllers. http://www.f5.com/pdf/white-papers/evolution-adc-wp.pdf Technology Primer: Transparent Application Delivery Networks. http://www.bluecoat.com/doc/5276 Bartlett, J.; Sevcik, P., How Network Transparency Affects Application Acceleration Deployment. http://www.riverbed.com/docs/AnalystReport- NetForecast-Transparency.pdf 11

Related