Transcript Deep Space Network Emulation

Deep Space Network Emulation
Shaun Endres and Behnam Malakooti
Case Western Reserve University
Department of Electrical Engineering and Computer Science
{Shaun.Endres, Behnam.Malakooti}@case.edu
Kul Bhasin and Allen Holtz
NASA Glenn Research Center
Space Communications
{Kul.B.Bhasin, Allen.P.Holtz}@grc.nasa.gov
Glenn Research Center
at Lewis Field
Introduction
Presentation Outline
•
•
•
•
•
•
•
•
Motivation
Application
Definition
Overview
Design
Configuration
Testing & Results
Conclusions
Glenn Research Center
at Lewis Field
Motivation
Network Analysis Motivation
• Performance assessment
• Risk assessment
Deep Space Emulation Motivation
• No known tools for evaluating deep space networks
– Space Based Internet Emulator (University of Kansas)
– The Network Simulator (ns-2)
• Active Networking
Glenn Research Center
at Lewis Field
Application
Deep Space Network Nodes (Earth, Moon, Mars, and beyond)
•
•
•
•
Ground stations
Satellites
Space shuttles
Anything that the Satellite Tool Kit (STK) can handle
Researchable Issues
• Protocols
• Routing algorithms
• Communication architectures
Glenn Research Center
at Lewis Field
Network Emulation
Definition
• Emulation refers to the ability to introduce a simulator into
a live network.
Comparison
• Simulation is _____ than emulation
– Less Expensive
– Faster (depending on the complexity)
– Less Accurate
• A real hardware test-bed is _____ than emulation
– Significantly more expensive
– More accurate
– Less controllable
Glenn Research Center
at Lewis Field
Emulator Overview
} User
} Emulator
Emulator Features
• Completely in user space
• Any network, transport, or application layer protocol possible
• Strict cutoff between emulator and user
Emulator Usage
• Emulator is transparent to the end user
Glenn Research Center
at Lewis Field
Emulator Design (Control & Emulation Network)
Control Network
•
•
•
•
•
Star Topology
Sends ‘Physical’ Information
TCP/IP Used
Link Always Available
Error Free Link
Emulation Network
•
•
•
•
Fully Connected Network
Sends ‘Networking’ Information
Researched Protocol Used
Link Availability Based on
‘Physical’ Information
• Error Prone Link
Glenn Research Center
at Lewis Field
Emulator Design (Information Flow)
Control Network
• Communicates with STK [7] to determine
position and link information
• Updates all nodes’ link tables periodically
Emulation Network
• Applies information contained in link
tables to all out going packets
Physical and Data Link Layers
• Modeled within the emulator code
Network, Transport, and Application Layers
• Actual implementation (left up to the
researcher to implement)
Glenn Research Center
at Lewis Field
Emulator (Physical Setup)
Hardware
• Cisco Catalyst 2900 series XL switch
• Intel Pentium III 600MHz processor
computer with 128MB of RAM
• Fast Ethernet
Software
• STK 5.0.4
• Microsoft Windows XP Professional
• Microsoft Visual Studio .NET 2002
Programming Language
• C#
Glenn Research Center
at Lewis Field
Results
Delay Verification
Delay (ms)
50
40
30
20
Measured
Target
10
0
1
2
3
4
5
6
7
8
9
10 11
Packet Number
Delay Accuracy gets better as the delay increases
• With a 30s delay, there is about a 0.1% error
Bit Error Rate Verification
• Bit error rate application has about a 0.25% error
Glenn Research Center
at Lewis Field
Conclusions and Future Work
Conclusions
• Deep Space Network Emulation is Possible
Without Kernel Modifications
• Great Flexibility Can Be Given to the User
• OS Interrupt Timer Is Limiting Factor
Future Work
•
•
•
•
Multiple Nodes
Bandwidth Restriction
Deep Space Mission Scenario Analysis
Protocol Testing for Deep Space Missions
Glenn Research Center
at Lewis Field
Thank You
Thank You for Your Time and Attention
Any Questions?
Glenn Research Center
at Lewis Field
Selected References
[1] http://www.isi.edu/nsnam/ns/ns-emulation.html
[2] Bateman, M., Allison, C., and Ruddle, A., “A Scenario Driven Emulator for Wireless, Fixed
and Ad Hoc Networks,” 2003.
[3] Zheng, P., and Ni, L. M., “EMPOWER: A Network Emulator for Wireline and Wireless
Networks,” IEEE INFOCOM, 2003.
[4] Zheng, P., and Ni, L. M., “EMPOWER: A Scalable Framework for Network Emulation.”
[5] Zheng, P., and Ni, L. M., “Experiences in Building a Scalable Distributed Network
Emulation System.”
[6] Ingham, D. and Parrington, G., “Delayline: A Wide-Area Network Emulation Tool.”
[7] www.stk.com
[8] Rallapalli, S., “Emulation of a Space Based Internet Communication Link: Design and
Implementation,” University of Kansas, 2000.
[9] Baliga, S. R., “Design of a Space Based Internet Emulation System,” University of Kansas,
2002.
[10] Endres, S., Griffith, M., Malakooti, B., Bhasin, K., and Holtz, A., “Space Based Internet
Network Emulation for Deep Space Mission Applications,” AIAA ICSSC, May, 2004.
[11] Sklar, B., “Digital Communications: Fundamentals and Applications,” Prentice Hall, 2001.
[12] Herrscher, D., and Rothermel, K., “A Dynamic Scenario Emulation Tool.”
[13] Carson, M., and Santay, D., “NIST Net – A Linux-based Network Emulation Tool.”
[14] Yeom, I., and Reddy, A. L., “ENDE: An End-to-end Network Delay Emulator.”
[15] Hu, Y., and Li, V. O. K., “Satellite-Based Internet: A Tutorial,” IEEE Communications
Magazine, 2001.
Glenn Research Center
at Lewis Field
Additional Slides
Additional Slides Begin Here
Glenn Research Center
at Lewis Field
Improvements to Existing Tools
Improvements to the Space Based Internet Emulator
•
•
•
•
Any transport and network layer protocol
No kernel modifications
Deep space objects
Header errors
Improvements to The Network Simulator (ns-2)
• Deep space objects
• Highly dynamic link changes
Glenn Research Center
at Lewis Field
Satellite Tool Kit (STK)
Satellite Orbit Propagation
• Detailed satellite propagation
• Orbits around any planetary body
Real time information
• STK/Connect can provide an application with
real time information
Link Budget Report
• STK/Communication can provide a detailed link
budget report
Glenn Research Center
at Lewis Field
Results
(1 of 2)
Worst Case Delay
Measured
1.00675
25.17%
10.0135
8
2.0135
25.17%
15.02025
12
3.02025
25.17%
15.02025
16
0.97975
6.12%
15.02025
20
4.97975
24.90%
25.03375
24
1.03375
4.31%
Measured
25.03375
28
2.96625
10.59%
Target
35.04725
32
3.04725
9.52%
35.04725
36
0.95275
2.65%
35.04725
40
4.95275
12.38%
45.06075
44
1.06075
2.41%
Delay (ms)
30
20
0
2
3
4
5
6
7
8
Error
4
40
1
Difference
5.00675
50
10
Target
9
10 11
Packet Number
Glenn Research Center
at Lewis Field
Results
Average Case Delay
Bit Error Rate
(2 of 2)
Measured
Target
Difference
Error
30030.49
30000
30.4865
0.10%
30050.51
30000
50.5135
0.17%
30035.49
30000
35.49325
0.12%
30035.49
30000
35.49325
0.12%
30020.47
30000
20.473
0.07%
30040.5
30000
40.5
0.14%
30020.47
30000
20.473
0.07%
30015.47
30000
15.46625
0.05%
30030.49
30000
30.4865
0.10%
30035.49
30000
35.49325
0.12%
Measured
Target
Difference
Error
1
1
0
0.00%
0.317118333
0.316228
0.000891
0.28%
0.099778333
0.1
0.000222
0.22%
0.03181
0.031623
0.000187
0.59%
0.010053333
0.01
5.33E-05
0.53%
0.003131667
0.003162
3.06E-05
0.97%
Glenn Research Center
at Lewis Field