1 / 66

A Switch Design for Real-Time Industrial Networks

This paper explores the design of a real-time switch for cyber-physical systems. It examines the need for real-time switches in various applications, such as telepresence, telemedicine, tele-robotic underground mining, and next-generation aircraft. The paper also discusses the challenges in designing a real-time switch and presents a solution compatible with the "crossbar + VOQ + iSLIP" architecture.

ccoates
Download Presentation

A Switch Design for Real-Time Industrial Networks

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. A Switch Design for Real-Time Industrial Networks Qixin Wang*, Sathish Gopalakrishnan**, Xue Liu***, and Lui Sha* * CS Dept., UIUC ** ECE Dept., UBC *** CS Dept., McGill Univ. 2/26/2008 Qixin Wang, UIUC,

  2. A Real-Time Switch Design for Cyber-Physical Systems Qixin Wang*, Sathish Gopalakrishnan**, Xue Liu***, and Lui Sha* * CS Dept., UIUC ** ECE Dept., UBC *** CS Dept., McGill Univ. 2/26/2008 Qixin Wang, UIUC,

  3. Many CPS applications need infrastructures built on real-time switches Real-Time Qixin Wang, UIUC,

  4. Telepresence lets people collaborate without moving to the same geographical location: needs RT Internet Qixin Wang, UIUC,

  5. Telemedicine, Tele-Surgery save more lives: need real-time Internet Qixin Wang, UIUC,

  6. Tele-robotic underground mining saves lives: needs real-time WAN > 5000 annual death toll 2000ft(609.6m) Qixin Wang, UIUC,

  7. Next Generation Aircraft: needs real-time high performance computer architecture, e.g., InfiniBand Concord 1976 Qixin Wang, UIUC,

  8. Next Generation Aircraft: needs real-time high performance computer architecture, e.g., InfiniBand A380 2007 Qixin Wang, UIUC,

  9. Next Generation Industrial Control: needs real-time fieldbus (plant area real-time control network) Qixin Wang, UIUC,

  10. Many CPS applications need infrastructures built on real-time switches Real-Time Qixin Wang, UIUC,

  11. Designing real-time switch is not easy • A. Demers, S. Keshav, and S. Shenker, “Analysis and simulation of a fair queueing algorithm,” in SIGCOMM’89, 1989. pp. 1-12. • D. Ferari, and D. Verma, “A scheme for real-time channel establishment in wide-area networks,” in IEEE JSAC, v8, n3, April, 1990 • D. C. Verma, H. Zhang, and D. Ferrari, “Delay jitter control for real-time communication in packet switching network,” in Proceedings of Tricomm’91, pp. 35-46, April 1991. • S. Golestani, “A framing strategy for congestion management,” in IEEE JSAC, v9, n7, pp. 1064-1077, Sep., 1991 • L. Zhang, "VirtualClock: a new traffic control algorithm for packet-switched networks," ACM Trans. on Computer Systems, Vol. 9, No. 2, May 1991 • A. K. Parekh, R. G. Gallager, “A Generalized Processor Sharing Approach to Flow Control in Integrated Services Networks: The Single-Node Case,” in IEEE/ACM ToN, v1, n3, June, 1993. • A. K. Parekh, R. G. Gallager, “A Generalized Processor Sharing Approach to Flow Control in Integrated Services Networks: The Multiple Node Case,” in IEEE/ACM ToN, v2, n2, April, 1994. • S. Jamaloddin Golestani, ``A self-clocked fair queuing scheme for broadband applications,'' Proc. IEEE INFOCOM'94, pages 636--646, IEEE, 1994. • H. Zhang, "Service Disciplines For Guaranteed Performance Service in Packet-Switching Networks," Proceedings of the IEEE, 83(10), Oct 1995 • Jon C. R. Bennett and H. Zhang, ``WF2Q: worst-case fair weighted fair queueing,'' IEEE INFOCOM'96, March, 1996 • Pawan Goyal, Harrick M. Vin, and Haichen Cheng, "Start-time fair queueing: a scheduling algorithm for integrated services packet switching networks," ACM SIGCOMM'96, pages 157--168, ACM press, August, 1996 • I. R. Philp, “Scheduling real-time messages in packet-switched networks,” Ph.D. Thesis, Technical Report No. UIUCDCS-R-96-1977, CS Dept., UIUC, Oct., 1996. • Subhash Suri, George Varghese, and Girish Chandranmenon, ``Leap forward virtual clock: A new fair queuing scheme with guaranteed delays and throughput fairness,'' IEEE Proc. INFOCOM'97, 1997 • I. Stoica, S. Shenker, and H. Zhang, ``Core-stateless fair queuing: achieving approximation fair bandwidth allocations in high speed networks,'' Proc. of ACM SIGCOMM'98, October 1998, pp. 118-130. • C. R. Kalmanek, H. Kanakia, and S. Keshav, “Rate controlled servers for very high speed networks,” in IEEE Global Telecommunications Conference, Dec., 1999. • W. Sun, and K. G. Shin, “End-to-end delay bounds for traffic aggregates under guaranteed-rate scheduling algorithms,” in IEEE ToN, 13(5): 1188-1201, Oct., 2005. Huge Volume of Literature since 1989 Qixin Wang, UIUC,

  12. Today’s Internet is still NOT real-time. Measured delay btw UK and Spain [Marsh06] Qixin Wang, UIUC,

  13. Today’s Internet is still NOT real-time. Why? Measured delay btw UK and Spain [Marsh06] Qixin Wang, UIUC,

  14. Industry acceptance has the final say. Two things rules: simplicity and switch hardware features Crossbar+InputQueueing Crossbar+VOQ+iSLIP HOLProblem OutputQueueing+QoS NSpeed-UpProblem Combined input-output queueing, ormore complicated switch hardware,to emulate output queueing+QoS (e.g. WFQ) Qixin Wang, UIUC,

  15. Past lessons tell us to design the real-time switch compatible to “crossbar + VOQ + iSLIP” Crossbar+InputQueueing Crossbar+VOQ+iSLIP HOLProblem OutputQueueing+QoS NSpeed-UpProblem Combined input-output queueing, ormore complicated switch hardware,to emulate output queueing +QoS (e.g., WFQ) Qixin Wang, UIUC,

  16. What is “Crossbar + VOQ + iSLIP”? A widely implemented switch architecture Simple High switch utilization Fast adaptation to random traffic(Internet traffic) Advantages Qixin Wang, UIUC,

  17. What is “Crossbar + VOQ + iSLIP”? Input Ports I1 I2 I3 Qixin Wang, UIUC,

  18. What is “Crossbar + VOQ + iSLIP”? Output Ports I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  19. What is “Crossbar + VOQ + iSLIP”? Virtual Output Queue (VOQ) I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  20. What is “Crossbar + VOQ + iSLIP”? cells I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  21. What is “Crossbar + VOQ + iSLIP”? I1 O1 O2 I2 cell cell cell cell cell O3 I3 cell Qixin Wang, UIUC,

  22. What is “Crossbar + VOQ + iSLIP”? Synchronous periodic cell forwardingCell-Time I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  23. What is “Crossbar + VOQ + iSLIP”? Matching I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  24. What is “Crossbar + VOQ + iSLIP”? Why Matching? An input/output can only send/receive one cell per cell-time I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  25. What is “Crossbar + VOQ + iSLIP”? I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  26. What is “Crossbar + VOQ + iSLIP”? I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  27. What is “Crossbar + VOQ + iSLIP”? I1 O1 I2 O2 I3 O3 Qixin Wang, UIUC,

  28. What is “Crossbar + VOQ + iSLIP”? iSLIP is like faculty job hunting  I1 O1 I2 O2 I3 O3 Request (apply job) Qixin Wang, UIUC,

  29. What is “Crossbar + VOQ + iSLIP”? iSLIP is like faculty job hunting  I1 O1 I2 O2 I3 O3 I1 O1 Request (apply job) I2 O2 I3 O3 Grant(offer job) Qixin Wang, UIUC,

  30. What is “Crossbar + VOQ + iSLIP”? iSLIP is like faculty job hunting  I1 O1 I1 O1 I2 O2 I2 O2 I3 O3 I3 O3 I1 O1 Request (apply job) Accept(take offer) I2 O2 I3 O3 Grant(offer job) Qixin Wang, UIUC,

  31. But we cannot directly adopt “crossbar + VOQ + iSLIP” Huge per hop delay bound E2E delay bound is an open problem [Gopalakrishnan06] Qixin Wang, UIUC,

  32. But we cannot directly adopt “crossbar + VOQ + iSLIP” Huge per hop delay bound E2E delay bound is an open problem [Gopalakrishnan06] What to do? Qixin Wang, UIUC,

  33. But we cannot directly adopt “crossbar + VOQ + iSLIP” Huge per hop delay bound E2E delay bound is an open problem [Gopalakrishnan06] What to do? Look at changed assumptions and new features. Qixin Wang, UIUC,

  34. Changed assumption: traffic predictability iSLIP is for non-real-time traffic: Unpredictable Qixin Wang, UIUC,

  35. Changed assumption: traffic predictability iSLIP is for non-real-time traffic: Unpredictable Flows change rapidly Unpredictable message size and arrival Qixin Wang, UIUC,

  36. Unpredictable traffic forces iSLIP to negotiate for each cell-time I1 O1 I1 O1 I2 O2 I2 O2 I3 O3 I3 O3 I1 O1 Request (apply job) Accept(take offer) I2 O2 I3 O3 Grant(offer job) Qixin Wang, UIUC,

  37. Just like faculty job hunting, iSLIP’s negotiate is dynamic, hard to give tight delay upper bound. Qixin Wang, UIUC,

  38. Changed assumption: traffic predictability iSLIP is for non-real-time traffic: Unpredictable Flows change rapidly Unpredictable message size and arrival RT-Switch for real-time traffic: Deterministic Qixin Wang, UIUC,

  39. Changed assumption: traffic predictability iSLIP is for non-real-time traffic: Unpredictable Flows change rapidly Unpredictable message size and arrival RT-Switch for real-time traffic: Deterministic Flows rarely change Regular message size and arrival Non-real-time traffic  real-time traffic Qixin Wang, UIUC,

  40. Deterministic traffic allows RT-Switch to forward cells like clockwork, no need to negotiate! Deterministic Grant is enough I1 O1 I2 O2 No need for Request and Accept I3 O3 Simplifies instead of extends iSLIP Qixin Wang, UIUC,

  41. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Qixin Wang, UIUC,

  42. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Small msg size (sensing: 1cell/msg, TV quality video: 480cell/msg) Qixin Wang, UIUC,

  43. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Small msg size (sensing: 1cell/msg, TV quality video: 480cell/msg) Short per msg transmission time (s: 0.5us, v: 240us) Qixin Wang, UIUC,

  44. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Small msg size (sensing: 1cell/msg, TV quality video: 480cell/msg) Short per msg transmission time (s: 0.5us, v: 240us) Large msg arrival period (s: 10ms, v: 30ms) Qixin Wang, UIUC,

  45. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Small msg size (sensing: 1cell/msg, TV quality video: 480cell/msg) Short per msg transmission time (s: 0.5us, v: 240us) Large msg arrival period (s: 10ms, v: 30ms) Reminds us of clock-driven time slicing OS Qixin Wang, UIUC,

  46. New feature: Large message arrival period allows clock-driven time slicing Real-time traffic Small msg size (sensing: 1cell/msg, TV quality video: 480cell/msg) Short per msg transmission time (s: 0.5us, v: 240us) Large msg arrival period (s: 10ms, v: 30ms) Reminds us of clock-driven time slicing OS RT-Switch runs fine time grain clock period (1ms), serving time slices (cell-time) to coarse time grain periodic RT flows Qixin Wang, UIUC,

  47. Solution: crossbar RT-Switch runs deterministic clock-driven time slicing Qixin Wang, UIUC,

  48. Solution: crossbar RT-Switch runs deterministic clock-driven time slicing Clock period of M cell-time, e.g., M = 5 Cell time: 1 2 3 4 5 I1: I2: I3: I4: Qixin Wang, UIUC,

  49. Solution: crossbar RT-Switch runs deterministic clock-driven time slicing Clock period of M cell-time, e.g., M = 5 Fit original real-time flow-forwarding tasks into clock period, e.g., (11, 3)  (5, 2), i.e., (10, 4) Cell time: 1 2 3 4 5 I1: I2: I3: I4: Qixin Wang, UIUC,

  50. Solution: crossbar RT-Switch runs deterministic clock-driven time slicing Clock period of M cell-time, e.g., M = 5 Fit original real-time tasks into clock period, e.g., (11, 3)  (5, 2), i.e., (10, 4) Demand Cell time: 1 2 3 4 5 a cell to send toO1 I1: a cell to send toO2 I2: a cell to send toO3 I3: a cell to send toO4 I4: Qixin Wang, UIUC,

More Related