A Methodology for Analyzing Complex Military    Command & Control (C2) Networks         For 10th ICCRTS, Track 4 (Assessme...
Purpose of PresentationA Methodology for Analyzing   Complex    Military      • Discuss the application of complex Command...
Introduction                  • Problem / Issue: Warfighter is faced with increasingly                    complex C2 netwo...
Trial of Network Analysis                                  MethodA Methodology for Analyzing    • Introduction of analysis...
Analysis Assumptions & Scope                  • Analysis AssumptionsA Methodology for Analyzing                     – A no...
Questions for AnalysisA Methodology     1.   Does the email cross domain solution change for Analyzing         previously ...
Network Metrics (Example)A Methodology                                            Link/node ratio = 1.33 (8 links, 6 nodes...
Network Metrics (Example)A Methodology for Analyzing   Complex                             Clustering coefficient    Milit...
Network Metric Thumb Rules                                        Experimentation and Analysis                         Met...
Question #1                   Does the email cross domain solution (CDS) change                     previously established...
Question #1                                   CDS Interactions                                Aggregate Network of UK Inte...
Question #2traffic flow?                                      Question #2                      Who are the key nodes for e...
Question #2                                                 Collaboration Measures                                    1000...
Question #3                    How robust isQuestion #3 light of the                                 the email network in ...
Question #3                                                           Robustness Measurement                              ...
Question #3                                               Degree Distribution                          Detailed Timeframe ...
Question #4                                  Question #4                   How does the structure of the email network evo...
Question #4                                              Number of Nodes                                                  ...
Question #4                              Link to Node Ratio (Distinct)       3.50       3.00       2.50kd/n       2.00    ...
Question #4                                     Network Progression                                                     Da...
Question #5                  What are the internal dynamics of select sub-networks and                       how to the su...
Question #5                                Network Diagram                   Staff #2 Sub-Network Interactions – Entire Ex...
Future Applications                  • Information Operations / Information AssuranceA Methodology        – Focus network ...
Next StepsA Methodology                  • Next experiment to use methodology for Analyzing   Complex                    C...
Questions?
Upcoming SlideShare
Loading in...5
×

A Methodology for Analyzing Complex Military Command & Control (C2) Networks

612

Published on

The results from a “discovery analysis” of military social networks used during a US/UK naval exercise. Presented to the 10th ICCRTS conference.

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
612
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
17
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

A Methodology for Analyzing Complex Military Command & Control (C2) Networks

  1. 1. A Methodology for Analyzing Complex Military Command & Control (C2) Networks For 10th ICCRTS, Track 4 (Assessment, Tools, and Metrics) Tysons Corner, Virginia June 14th, 2005 By: David A. Jarvis Originally prepared under contract for Navy Warfare Development Command (NWDC) ©2005 Alidade Incorporated. All Rights Reserved
  2. 2. Purpose of PresentationA Methodology for Analyzing Complex Military • Discuss the application of complex Command & Control (C2) network theory principles to military C2 Networks networks • Present results of CJTFEX 04-2 analysis • Summarize validity and potential applications of C2 Network Analysis (C2NA) method 10th ICCRTSJune 14th, 2005
  3. 3. Introduction • Problem / Issue: Warfighter is faced with increasingly complex C2 networksA Methodology for Analyzing – Increasing number of IP networks, communication networks, and Complex applications all creating a complex information environment Military – Warfighter’s capability and effectiveness of new applications and Command & networks are difficult to analyze Control (C2) Networks – Traditional C2 analyses limited to IT performance and human interface • Possible Solution – New analysis techniques can now be applied to define the structure, dynamics and evolution of collaboration in command and control networks • A “network” is any collection of interacting elements arranged for purpose, not necessarily an IT network – Techniques enable the analysis of how warfighters actually use networks, as opposed to how engineers tell us how to build them 10th ICCRTS – Metrics can be used in defining and measuring new informationJune 14th, 2005 architectures
  4. 4. Trial of Network Analysis MethodA Methodology for Analyzing • Introduction of analysis method within CJTFEX Complex 04-2 (12-day joint US/UK exercise) Military Command & – Existing Cross Domain Solution (CDS) Limited Control (C2) Objective Experiment data collection used to validate Networks method – Performed in addition to traditional NWDC Analysis effort • Analysis Focus - Email – The analysis is applicable to a wide range of networks, email used as a stepping stone – Email is the primary method of asynchronous electronic communication in the Information Age – Indicates structures of collaboration and command 10th ICCRTS and controlJune 14th, 2005
  5. 5. Analysis Assumptions & Scope • Analysis AssumptionsA Methodology for Analyzing – A node is an email address Complex – A link exists if at least one email has been exchanged Military Command & between two nodes, TO: and CC: are treated Control (C2) identically Networks – Structural analysis only, the content or intent of messages not considered – Artifacts exist in the raw data (e.g. record message traffic), corrected where possible • Analytical Scope – Overview Analysis of six hour timeframes (based on battle rhythm), analyzing each independently 10th ICCRTS – Detailed Analysis on two selected timeframes,June 14th, 2005 demonstrating structure and dynamics of C2 network
  6. 6. Questions for AnalysisA Methodology 1. Does the email cross domain solution change for Analyzing previously established operating procedures? Complex Military 2. Who are the key nodes for email traffic flow? Command & Control (C2) Networks 3. How robust is the email network in light of the removal of nodes and/or links? 4. How does the structure of the email network evolve over the course of the experiment? 5. What are the internal dynamics of select sub- networks and how to the sub-networks interact with each other? 10th ICCRTS These are the questions…June 14th, 2005 what metrics provide the answers?
  7. 7. Network Metrics (Example)A Methodology Link/node ratio = 1.33 (8 links, 6 nodes) for Analyzing Complex Degree distribution (histogram) 3 Military Degree Command & # of nodes 2 Control (C2) 2 D C 1 Networks 1 4 3 B E 0 1 2 3 4 5 6 7 8 # of connections per node A Characteristic path length (CPL) F 2 4 A B C D E F Avg A 0 1 2 2 2 1 1.6 Node Link B 1 0 1 2 1 1 1.2 C 2 1 0 3 2 2 2.0 D 2 2 3 0 1 1 1.8 E 2 1 2 1 0 1 1.4 10th ICCRTS F 1 1 2 1 1 0 1.2June 14th, 2005 CPL = 1.5 (median of the averages)
  8. 8. Network Metrics (Example)A Methodology for Analyzing Complex Clustering coefficient Military 3 x # of triangles Command & C D C= Number of connected triples of nodes Control (C2) Networks C = 3 x 3 / 34 = ~0.26 B E Betweenness - A measure of network resilience and flow A F g jk ( ni ) C B ( ni ) = ∑ j<k g jk C B ( ni ) = Betweenness centrality for node i g jk = # of shortest paths (geodesics) linking the two actors j and k g jk ( ni ) = # of shortest paths (geodesics) linking actors j and k that contain i 10th ICCRTSJune 14th, 2005
  9. 9. Network Metric Thumb Rules Experimentation and Analysis Metric Range Operational SignificanceA Methodology Number of nodes, n n > ~100 Network effects unlikely to occur with n < 50 for Analyzing Number of links, l l < ~2n l << 2n, too brittle Complex l >> 2n, too much overhead Military Command & Degree distribution Skewed Adaptivity, modularity Control (C2) Largest hub < 100 links Hub appears, recedes by reconnection 5% of links Networks Average path length log(n) Short distances even for large networks (e.g., 104 nodes Average path length = ~4) Clustering Skewed Hierarchy, organization Betweenness Skewed Cascade control Path horizon log(n) Self-synchronization Susceptibility/ Low (random removal) Hubs should be kept obscure until needed, damage Robustness High (focused removal) abatement/repair schemes Neutrality (0, 2) Increased network effects, decreased susceptibility, tipping points, max = n/2 10th ICCRTSJune 14th, 2005 Metrics measure how people interact in a military context ©2005 Alidade Incorporated. All Rights Reserved
  10. 10. Question #1 Does the email cross domain solution (CDS) change previously established operating procedures?A Methodology for Analyzing Complex Military • We found: Command & Control (C2) – CDS increased integration between US and Networks UK networks – Additional baseline information required to fully define cross domain email need and use • Method supports: – Defining role for individual liaison officers 10th ICCRTSJune 14th, 2005
  11. 11. Question #1 CDS Interactions Aggregate Network of UK InteractionsA Methodology = UK for Analyzing Complex = US Military Command & Control (C2) Networks 10th ICCRTSJune 14th, 2005 Multiple conduits between domains
  12. 12. Question #2traffic flow? Question #2 Who are the key nodes for emailA Methodology • Based on multiple metrics, we found: for Analyzing Complex – J2 ACOS Military Command & – Information Operations Control (C2) Networks – Asst. JOC Watch • Method supports: – Developing network defense for most important nodes – Providing input to plans for graceful degradation of capability – Examining use of method to exploit adversary 10th ICCRTS networks and C2 structureJune 14th, 2005
  13. 13. Question #2 Collaboration Measures 1000A Methodology for Analyzing Complex Military Receivers s Command & 100 at or In-Degree (kin) Control (C2) Networks b or a C oll 10 Broadcasters 1 1 10 100 1000 10000 Out-Degree (kout) Timeframe Receive Only Xmit Only Xmit & Receive 10th ICCRTS Day 6 1200-1800 684 (56%) 91 (7%) 441 (36%)June 14th, 2005 Day 8 1200-1800 894 (58%) 146 (10%) 504 (33%)
  14. 14. Question #3 How robust isQuestion #3 light of the the email network in removal of nodes and/or links?A Methodology for Analyzing Complex • We found: Military Command & – Resilient to random node removal Control (C2) Networks – Vulnerable to targeted node removal – Network structure makes rapid recovery possible • Method supports: – Critical node placement in distribution of staff – Development of alternate C2 paths 10th ICCRTS – Improving node counter-targetingJune 14th, 2005
  15. 15. Question #3 Robustness Measurement Detailed Timeframe – Day 8 1200-1800A Methodology 1600 for Analyzing Complex 1550 Military 1500 Command & 1450 Control (C2) Size of Giant Component Networks 1400 1350 1300 Degradation is not linear 1250 1200 1150 1100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Number of Nodes Deleted Targeted Random 10th ICCRTS In targeted case ~2% of nodes removed, ~25% of network lostJune 14th, 2005
  16. 16. Question #3 Degree Distribution Detailed Timeframe - Day 8 1200-1800 1000A Methodology for Analyzing y = 170.52 * x^(-1.2346) R= 0.98488 Complex y = 625.1 * x^(-1.4355) R= 0.9956 Military Command & Control (C2) 100 Networks FREQUENCY 10 OUT DEGREE (<101) IN DEGREE (<118) 1 1 10 100 DEGREE 10th ICCRTSJune 14th, 2005 Skewed distribution is evidence of a scale-free network
  17. 17. Question #4 Question #4 How does the structure of the email network evolve over the course of the experiment?A Methodology for Analyzing • We found: Complex Military – Network structure follows staff daily battle Command & Control (C2) rhythm, significant events did not alter the Networks network structure – Distance to get information from one person to another remained roughly constant • Method supports: – Re-engineering networks based on user behaviors to assist in meeting warfighter 10th ICCRTS requirementsJune 14th, 2005
  18. 18. Question #4 Number of Nodes Number of Links 2500 18000 16000 2000 14000 12000 1500 10000 k, kdn 8000 1000 6000 4000 500 2000 0 0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 Day 11 Day 12 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 Day 11 Day 12 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 # of total links # of distinct links Clustering Coefficient 0.30 • Network size cycles with daily 0.25 battle rhythm 0.20 • Link activity cycles with daily 0.15 battle rhythmC 0.10 • Clustering coefficient has no obvious pattern 0.05 0.00 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 Day 11 Day 12 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400 2400
  19. 19. Question #4 Link to Node Ratio (Distinct) 3.50 3.00 2.50kd/n 2.00 1.50 Number of Nodes vs. Number of Distinct Links 2500 1.00 2000 0.50 1500 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 Day 11 Day 12 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 1800- 2400 2400 2400 2400 2400 2400 2400 2400 n 2400 1000 2400 2400 2400 500 • Ratio approximates thumb rule • Link activity grows linearly 0 0 1000 2000 3000 4000 5000 6000 7000 kd with addition of nodes
  20. 20. Question #4 Network Progression Day 5A Methodology for Analyzing Complex Military Command & Control (C2) Networks = US & UK Time (EDT) n k kd k/n kd/n CPL log(n) C 0000-0600 557 2636 984 4.73 1.77 5.02 2.99 0.08 0600-1200 2069 16,700 5558 8.07 2.69 4.78 3.74 0.05 10th ICCRTS 1200-1800 1449 13,209 4329 9.12 2.99 4.52 3.64 0.12June 14th, 2005 1800-2400 1183 8867 2928 7.50 2.48 4.83 3.47 0.20
  21. 21. Question #5 What are the internal dynamics of select sub-networks and how to the sub-networks interact with each other?A Methodology for Analyzing • We found: Complex Military – Structures of the sub-networks were very different Command & Control (C2) from entire CJTFEX email network, the CJTFEX Networks was scale-free, the staff sub-networks were not – Identifiable nucleus of communications in each staff – The two nuclei of Staff #1 and Staff #2 were well- connected – Using different link definitions (reciprocal, threshold) can provide additional information about the network • Method supports: 10th ICCRTS – Development of techniques to split staffs betweenJune 14th, 2005 assets
  22. 22. Question #5 Network Diagram Staff #2 Sub-Network Interactions – Entire Exp. (Reciprocal Link Definition)A Methodology for Analyzing Complex Military Command & Control (C2) Networks = nucleus node 10th ICCRTSJune 14th, 2005
  23. 23. Future Applications • Information Operations / Information AssuranceA Methodology – Focus network defense on most important nodes for Analyzing Complex – Improve node counter-targeting Military – Examine use of method to exploit adversary networks and C2 Command & structure Control (C2) Networks • C2 Structure and Information Flow – Support decision of critical nodes placement in distribution of staff – Develop alternate C2 paths – Measure and understand key command and staff relationships to more effectively use Collaborative Information Environment • Network and Information Management – Assist warfighter in defining requirements and providing feedback on engineering design parameters – Provides metrics for evaluation and design of information 10th ICCRTS management practicesJune 14th, 2005 – Provide input to plans for graceful degradation of capability
  24. 24. Next StepsA Methodology • Next experiment to use methodology for Analyzing Complex CDS LOE II (JTFEX 05-2) Military Command & • Multi-level C2 analysis, combining Control (C2) Networks multiple C2 systems – Email – Chat – Voice Over IP • Content analysis • Incorporation of lessons learned from 10th ICCRTS CDS LOE I (JTFEX 04-2)June 14th, 2005
  25. 25. Questions?
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×