An Open Strategy for the Acquisition of Models and Simulations

Transcription

1 Calhoun: The NPS Institutional Archive Reports and Technical Reports All Technical Reports Collection An Open Strategy for the Acquisition of Models and Simulations Rudy Darken

2 An Open Strategy for the Acquisition of Models and Simulations Rudolph P. Darken Director, MOVES Institute 1

3 Focus Not about the use of simulation for acquisition but rather the acquisition of simulation itself. Estimates of DoD M&S acquisition expenditures range from $9B to $15B annually. Almost exclusively a proprietary software industry segment. Why? Is this good for DoD? Can we do better? 2

4 What do these have in common? A lot actually rendering, devices, networking, etc. 3

5 Current practice First Application Second Application HLA Bus M&S Application M&S Application Vendor A Vis Sim Engine A OS Hardware Vis Sim Engine B OS Hardware Vendor B Application Level Reusability = 0 4

6 A (slightly) better way First Application Second Application HLA Bus M&S Application M&S Application Vendor A Vis Sim Engine X OS Hardware Vis Sim Engine X OS Hardware Vendor B Reused code 5

7 A (slightly) even better way First Application Second Application HLA Bus M&S Application M&S Application Vendor A Vis Sim Engine X OS Hardware Vis Sim Engine X OS Hardware Vendor B Government owned source Reused code 6

8 Still not so great First Application Second Application HLA Bus Vendor A M&S Application Vis Sim Engine X OS Hardware Government owned source Reused code M&S Application Vis Sim Engine X OS Hardware Recurring license fees Locked in to Engine X Vendor B 7

9 Is open source a candidate solution? Recommends the adoption (or full consideration) of OSS in DoD software acquisition. Still lacking in policy about government contribution to OSS. 8

10 Available funding Why OSS? Service, Maintenance, Upgrades Applications Proprietary features Core functions Available funding Service, Maintenance, Upgrades Applications Proprietary features Core functions 9

11 Four design tenets 1. Keep everything open to avoid lock-ins and increase flexibility 2. Make it multi-genre since we never know what app it's going to have to support next 3. Make it modular so we can swap anything out as technologies mature at different rates 4. Build a community (or leverage existing ones) so the military doesn't have to pay all the bills 10

12 Detailed breakdown Over 90% of the source code was written by someone else, not paid for by the government. 11

13 Full development environment 12

14 13

15 Did we meet our goals? almost, but not quite. 14

16 Four design tenets 1. Keep everything open to avoid lock-ins and increase flexibility 2. Make it multi-genre since we never know what app it's going to have to support next? 3. Make it modular so we can swap anything out as technologies mature at different rates 4. Build a community (or leverage existing ones) so the military doesn't have to pay all the bills 15

17 An even better way Service Oriented Architecture Business logic Function Application Business logic Function Web services, published standard APIs Business logic Function Function Data Data Business logic Enterprise Service Bus (ESB) Data Data From monolithic to distributed. 16

18 Why SOA? Way more flexible Faster implementation (once infrastructure is in place) Maximize reuse Create a MARKET for your business Will never work (in its literal form) for real-time simulation Web services are too slow but I want all this stuff 17

19 A real time SOA for simulation HLA Bus Application A data Application B data Data driven simulation Render Network AI UI Runtime OS Game Logic Render Network AI UI Runtime OS Game Logic Simulation Service Bus Published APIs Hardware Hardware Government owned Open source Proprietary 18

20 A test case for USMC infantry training Urban combat gamebased trainer DVTE Same application! Urban combat small weapons trainer ISMT Urban combat small team trainer SITE 19

21 Take aways Open source has a place in Defense Modeling and Simulation Open source encourages vendor innovation by concentrating effort on what s new DoD must architect simulation frameworks to create a marketplace for our business This helps build the commercial defense simulation industry Reusability is key We want what an SOA offers but we need it in real time. 20

22 Thank you for your attention Rudy Darken Director, MOVES Institute Naval Postgraduate School