RE: UPIX data analysis

From: Tabish Hasan <tabish_at_mediadefender.com>
Date: Mon, 9 Apr 2007 17:33:59 -0700

For our interdiction presence test, we tested the Black Donnellys on
torrentz.com. Dan and Jon grabbed all the hashes from torrentz and
checked whether or not that torrent is in our inderdiction system.
Currently, all of them ARE within our system. However, when we did our
first check, many of the torrents were NOT in our system because they
got moved into the dead_torrents folder. After we moved the torrents
back into the protection_enabled folder, we now have 100% coverage.
Every torrent w/in torrentz.com is either our decoy or is within the
interdiction system.

 

Also, our interdiction system is handling well. All the DL tests that we
started are still stuck at less than 10% with no signs of them finishing
anytime soon. We're currently at 800 or so torrents.

 

Below is every Black Donnellys hash on torrentz.com and the
corresponding status.

 

Hash

Real / Decoy

Interdiction

b136d9e642b4d7660bdfcebd3c5799de9b2022d0

Real

Active

8af2d70e75c19a72e337f231478151b0cabe94f4

Real

Active

d5792f3ac6d138a31e751cd9632c6bce5c17cc60

Real

Active

5e790d293db1764d66dcf9e88d2438b785b50f29

Old-Skool Decoy

 

09590b38391ed3529d4a1d54c8b0d9b0f87c03d8

Regular Decoy

 

fcc066ed1988217fb34fb741661ff3517e916b98

Real

Active

6f29472e19fdc027470fc256153438013edf4e94

Old-Skool Decoy

 

cb24d586535c2464acf1a163322c04fda25c388f

Real

Active

04435dc2b3f6074c388ee5ae54c4a8e4619e496d

Real

Active

80781a830419618bdb49e42538229f62e126783d

Real

Active

a35cc17b2ad2ffcbf63fdb7dff82931d63b7eb3f

Regular Decoy

 

53591bc44c74f8c19cfcd5af832cc235be471d5d

Old-Skool Decoy

 

b911cc2f462b8b84abca03e5524e4421f12c7ae2

Old-Skool Decoy

 

37e120f73eddb42d29a74302c492768aa097cd8b

Real

Active

e9f93791081a00decc44c705d6b13da2d2c945b8

Real

Active

2056b2b2414d2d1b7fdf37caad9ca02dce545d38

Regular Decoy

 

fbc3945e230bb066b7533e247544fd6e23bdd116

Real

Active

ea147917a1502e8d8119843c218a64444665496a

Regular Decoy

 

84809084cdfc89029dbd13df2e8e113900a1b9d8

Real

Active

54e02f154566267a9e0e00663673bf3a37ac0cd7

Real

Active

6ede13274a81c8cd1ca8e649a25c9528751bfde6

Real

Active

7a88108bae31ec83b1b7abe96cf38bacea793555

Old-Skool Decoy

 

97fe939230e046312f1b2a5632e68f182c6261e4

Real

Active

e4fa2b2f865952976996b0219ad258d549a617f9

Real

Active

5aab928c509c725460d366026eabb33d7e59ca7d

Real

Active

18f866e51988e23a801ec418c0e98fe51a05dc3f

Real

Active

106bf43ec324e072bf41c4cf19f87a31c9276be8

Real

Active

7dd4af7ba37d61e6db8d7a91eefdb5ef42fa2da9

Real

Active

2f98f3369415a74775f14234713c0470b2eab9be

Real

Active

c24d37a2be6174aceb96b5a40f8b440eb6b42999

Old-Skool Decoy

 

6987994a12870d628af193dbb42c687de152f764

Real

Active

19142adf4f59c86ee460db7a8d1521ecd4469367

Old-Skool Decoy

 

c6a65da8254d2a3e9596a407b2af1ef376a109ea

Real

Active

2fb70365a9cbd25df122277675699a94fc9c11df

Real

Active

15d47a51d15864bd063d174a7c1cbe09923c7a0f

Real

Active

41529ea75b1eeedf8c71f415aa069e4e59581594

Old-Skool Decoy

 

d252932a878ff1e94d3e5f6e8f52e118972e546b

Real

Active

a576d2ded1ac864ef3e5abcbde9f201e6b8199d5

Regular Decoy

 

044b5f6ab1e7fbb65d1d255faa409e7c07d7e67d

Real

Active

93136c231f25b7f6536625e98ba8d0df6a5cfbc9

Real

Active

0447fe7e394c26e780e430f9d2abfc1d42cd4ef3

Real

Active

e7d6a1493366669f44c470a25b2fed5737b65e7f

Real

Active

648ba20c47e8e50d13b6a2112ff3936f936ebc76

Old-Skool Decoy

 

dd60bff864d7cbb3a9c0d79fff3aca57139318ea

Old-Skool Decoy

 

6524efdc2917abf1ff9ab9e0e57bd9ca4cef5a0b

Regular Decoy

 

1cf5dc0e48582c6648ff2eb5d8622f2ebd30c37a

Real

Active

0254ce1183d2e7b4e311d22cad8129953f620e19

Real

Active

ccc2f190e7f7146a72ebdfabbdf1b95ae68bdac4

Regular Decoy

 

6472d2761d164d85c519891a2241cc43b3d3a26a

Real

Active

8b86772d01ac8b8e3099aa64a5a6ac2694d75965

Real

Active

950a1bde057011262c8bd6c3d6ca571d4b7cdb4a

Real

Active

b4db4ed22020f8b2ccf4f31d272563d801ae8225

Regular Decoy

 

11c28589659cca289c544b0dfe48b3a835ee6c1f

Real

Active

109806e0f1be12218f9fc86870f8a9e5c053ed43

Real

Active

927a83c7f7be5459a16f7223532f17b7647fd435

Real

Active

f46fe280186e138bccd2d3a44db466dcdd3e0305

Real

Active

bc612eeb7751c92a60dd2f2593e1dd52234a4421

Real

Active

7f9df00e499e975365b8f5b1f45264d5a858a4b1

Real

Active

2e4488c172321e8c5bbf2aa02b4bfd1d6c40d534

Regular Decoy

 

e7dda9522625e65a73b0f250a1ca82956ffe2dd0

Real

Active

149385ce8a46a0246ad65af00daa38ca2be80b83

Regular Decoy

 

cf8fe712e3cdcf32116aec385b6c394e2cd0f2c4

Real

Active

6df21814eb3b7043b63eb35d627f91af815cb31b

Real

Active

638aefcc1a03a9d2af3291a4c54bdff0f53c5abc

Real

Active

a43ff0b27180b438b85865a88bc4eebf21147fc8

Real

Active

5b8f32090e3a6b00f7fd5e2523393267d3158d9b

Real

Active

1b3e1d24bf731be215ed4d563247eb244da936a2

Real

Active

959092058bcdd340fdd53cef947a2ddea69703ac

Real

Active

743133e46b2fab29cbfb0785c42bb76960bfa306

Real

Active

2151ac5a51a619028060af3e51462b449ad34e6e

Regular Decoy

 

fc67e1a9a698e6df0a474c72c5a2c3320ed10acf

Real

Active

a242a57b4a0f28563a3162a28d4c17a699fbe8ae

Real

Active

 

________________________________

From: Tabish Hasan
Sent: Friday, April 06, 2007 6:20 PM
To: Ben Ebert; Randy Saaf; Jay Mairs
Cc: Ben Grodsky; Ty Heath; Daniel Lee; Jonathan Perez
Subject: RE: UPIX data analysis

 

We're running two types of interdiction tests on Monday. One type is
where we do actual DL testing and see how our interdiction system is
handling (with the addition of the torrents Jed found). And we will also
do an interdiction presence test, to see out of all the Black Donnellys
torrents on torrentz.com, what percent actually made it to our system.

 

Also, we have interdiction tests running today, so we can compare that
with Monday's data.

 

________________________________

From: Ben Ebert
Sent: Friday, April 06, 2007 6:00 PM
To: Tabish Hasan; Randy Saaf; Jay Mairs
Cc: Ben Grodsky; Ty Heath
Subject: RE: UPIX data analysis

 

The first week happens to be the exact same day as the QA team tested
interdiction on black donnellys 3/19. Our internal test was 100%
effective as well, that was the first week we saw really good results.
I do not know what happened the second week, we don't have a recent
interdiction test. It is possible from the way the data looks that they
tested exclusively from mininova the first week, and from snarf-it the
second week, and we were not hitting snarf-it. We will need to be
running interdiction tests on Monday.

 

________________________________

From: Tabish Hasan
Sent: Friday, April 06, 2007 5:47 PM
To: Randy Saaf; Jay Mairs
Cc: Ben Ebert; Ben Grodsky; Ty Heath
Subject: UPIX data analysis

 

For their 2 week data:

* Black Donnelly's on BT actually produced better results on week
1 than week 2. On the attached BT sheet, I've highlighted in yellow all
the ones that were good downloads. Notice that they all came in week 2.
Week 1 was 100% effective. Week 2 was 30% effective. Without knowing
the infohash, I can't determine if we had interdiction active for all
the torrents they attempted.
* There ARE several instances where they grabbed CD1 of a file,
but no corresponding CD2. Those are highlighted in red on the attached
eDonkey sheet.
* They're NOT attempting only the most recent season. An example
is their "Office", "House MD", and "Law and Order" tests below.

NBC - The Office - 205 - Halloween.avi

NBC - The Office - 209 - Party.avi

NBC - The Office - 302 - The Convention.avi

The Office (NBC) - 214 - The Carpet.avi

 

House MD - 102 - Paternity - hdtv-lol.avi

House MD - 2x03 - Humpty, Dumpty.avi

House Md - 2x10 - Failure To Communicate Hdtv Xvid-Fov.avi

House MD 2x13 - Skin Deep.avi

 

law and order svu 220 dvdrip.avi

law and order svu 504 loss hdtv-lol.avi

law and order svu 804.avi

law and order svu 8x14.avi

law and order svu s07e20 fat -.avi

law and order svu s08e06 infiltrated hdtv xvid-notv.avi

* They didn't test on Fastweb (Italian server) during this
period.

 

-TH
Received on Fri Sep 14 2007 - 10:56:04 BST

This archive was generated by hypermail 2.2.0 : Sun Sep 16 2007 - 22:19:48 BST