[DEECo] demo & analysis

[Jaroslav Keznikl]

Hi
if you're interested, you can try to analyze the demo. I have managed to
get some interesting results; see: http://pastebin.com/J3fqYhT6

I have used the configuration in ...-demo/configurations/...
I.e., teams are located to a single cluster.

You can try running analysis/analyze_log.py and analysis/analyze_demo.py on
the produced log file.


Excerpt of interesting results:

*Generic stats-------------------------------------------*
Average time per hop:  133.928970197 ms
Max number of hops:  9
Average number of hops:  2.17774628532
Boundary prevented sending of: 6623
KnowledgeData published: 10720
KnowledgeData received: 49355
Sent messages:  10940
Received message Ids:  5372
Dropped message Ids:  4932
Received/Sent ratio: 0.491042047532
Message size: avg=2926.873857, min=2000.000000, max=3000.000000,
median=3000.000000
Rebroadcasts planned:  16663
Rebroadcasts aborted:  7543
Rebroadcasts finished:  8740
Average abort ratio:  0.45267958951

*Scenario-specific stats:-------------------------------*
Process-2-process response time: avg=2914.765306, min=221.000000,
max=8335.000000, median=2553.500000
Node-2-node response time: avg=2403.204082, min=19.000000, max=7845.000000,
median=2064.000000
Hops: avg=2.581633, min=1.000000, max=6.000000, median=2.000000
Process-2-process response time / 1 hop: avg=1327.678571, min=169.000000,
max=6121.000000, median=1122.000000
Node-2-node response time / 1 hop: avg=1058.096939, min=19.000000,
max=5036.000000, median=854.000000
Version difference: avg=1.887755, min=0.000000, max=7.000000,
median=2.000000
Ratio of discovered vs. in danger: 196 of 221 (88.687783%)

I hope this is close to acceptable.

J
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://d3s.mff.cuni.cz/pipermail/deeco/attachments/20140208/8b6d124d/attachment.html>