Presentation is loading. Please wait.

Presentation is loading. Please wait.

My Experiences, results and remarks to TCP BW and CT Measurements Tools Jiří Navrátil SLAC.

Similar presentations


Presentation on theme: "My Experiences, results and remarks to TCP BW and CT Measurements Tools Jiří Navrátil SLAC."— Presentation transcript:

1 My Experiences, results and remarks to TCP BW and CT Measurements Tools
Jiří Navrátil SLAC

2 My interests concerns on
How to achieve full BW utilization Analysis ofTools for BW estimation New methods of CT detection

3 Inspiration Les Cottrell IEPM metrics Jin Gujun (LBNL) netest-2 recommendation Feng Wu-Chun (LANL) simulations Diana Katabi, Charles Blake (MIT) bottleneck detection

4 Area of saturation = area of full utilization
TCP windows size (not sufficient for HSL) Parallel streams combination

5

6 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit
Saturation point

7 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

8 No big difference between speed for P2,P4,P8, ..
Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit No big difference between speed for P2,P4,P8, ..

9 Inflexion point Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

10 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

11 Why such distribution of throughput for individual streams ?

12 Internet path Out In Virtual queue

13 Problems of slow and high loaded lines
All parallel streams share same “virtual queue” All “my traffic” share same queue with outside CT visible in statistics for “streams speed” visible in time reports (In iperf)

14 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

15 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

16 Why I cannot achieve more bandwidth with less streams?
Is there a CT ? I don’t know !

17 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

18 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

19 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

20 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

21 The other TCP applications behave very similarly in same environment
Conclusion: The other TCP applications behave very similarly in same environment But not all applications are so aggressive as Iperf !

22 Used tools for BW estimations
Pathrate Pathload Iperf Netest-2 Incite BWe UDPmon

23 Short characteristic of methods
Patharate + accurate - not very reliable, long run time Pathload + accurate, fast, light - limited range of operation ( < 155 Mbps)

24

25

26 Small comparisons

27 Short characteristic of methods
Iperf + reliable - must be configured to use full BW and postprocessed - heavy load of lines during operation Netest-2 + reliable, good reports - must be configured to use full BW, - accurate (different timing scheme)

28 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

29 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

30 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

31 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

32 INCITE: Edge-based Traffic Processing and Service Inference for High-Performance Networks
Richard Baraniuk, Rice University; Les Cottrell, SLAC; Wu-chun Feng, LANL Incite BW estimator

33 P2 P1 BWe ~f(VQ) =f(dTR) dTR <0,RTT> dTS (20 ms) Time P1 P2 P1 P2 dtp2 dTR RTT P1 P2 dtp1

34 Plop – linux (LANL) 2x 22 Mbit, 4x24 Mbit, 8x24Mbit

35

36

37 dt ~ F(VQ) ~ F(Internet)
freq= volume/dt volume ~ pkt_length*8 dt ~ F(VQ) ~ F(Internet) BWe=Mean(freq) Open Problem What is pkt_length?

38 Virtual queue All “my test traffic” share same queue with outside CT
(It means that delay caused to my pkts by VQ is not dependent only on my pkt_lengths !) The accuracy is dependent on knowlegde of packet distribution on particular path. The average packet length ~1000 bytes gives reasonable results. Virtual queue

39

40

41

42

43

44

45

46

47

48

49

50

51 Filtering & Averaging ANL Current/Average CERN Current/Average

52

53

54 Good Morning !

55 KPNQest UK-problems

56 Load balancing on one path router

57 Incite CT detection CT = f (dTR)
INCITE: Edge-based Traffic Processing and Service Inference for High-Performance Networks Richard Baraniuk, Rice University; Les Cottrell, SLAC; Wu-chun Feng, LANL Incite CT detection CT = f (dTR)

58

59 MF-CT Features and benefits
No need access to routers ! Current monitoring systems for Load of traffic are based on SNMP or Flows (needs access to routers) Low cost: Allows permanent monitoring (20 pkts/sec ~ overhead 10 Kbytes/sec) Can be used as data provider for ABW prediction (ABW=BW-CT) We have 2 data points of CT per second !! Weak point for common use MATLAB code

60

61

62

63

64

65

66

67

68

69 MF-CT verification model
UDP echo SNMP counter CERN SNMP counter internet SNMP counter SNMP counter IN2P3 SLAC SNMP counter MF-CT Simulator UDP echo


Download ppt "My Experiences, results and remarks to TCP BW and CT Measurements Tools Jiří Navrátil SLAC."

Similar presentations


Ads by Google