DEFCON 26 BRIEFING

Busy year this year! Unfortunately, several of our duplicators (about 40 duplication slots) didn't make it to Vegas this year.  We were hoping to get more duplicators so we took in 473 drives for duplication before we closed the intake.  We used all 84 of our available duplication slots for every available second from Thursday afternoon until Sunday morning to make all the data!

 

 

General Intake

20180809212653-6df33fda-me.jpg

Many thanks go out to all our volunteers this year - especially the new ones! We processed 319 drives and (again) cleared out all of the available 6TB drives in Vegas this year (you're welcome, Fry's). This is the breakdown of what requests we took in within the first hour at an average rate of 5.4 per minute on Thursday afternoon:

  • infocon.org archive: 99
  • Hash Tables 1 of 2:  121
  • Hash Tables 2 of 2: 99
 

DDV Drive content requests

 

As much as it pained us to do, we had to return an additional 145 drives back to their owners without content. Some people donated their drives to the DDV which we will put to good use next year.

 


DRIVE DETAILS

This year we captured even more drive details. We added a quick procedure that allowed us to catalog the brand and model of each drive in about a second.  Using this process after accepting a drive but before processing, we were able to capture information for all 319 drives we processed plus a few we didn't.  This gave us stats on a total of 363 drives.  We've dropped that detail for you below:

 

Intake by Manufacturer

 
 
BRAND      MODEL                QTY   FAILS  % FAIL
--------   ------------------   ----  -----  ------
Toshiba    X300                  55     0      0.0%
           N300                  21     0      0.0%
Seagate    Barracuda (dsktop)    29     4     13.8%
           Barracuda Pro         15     0      0.0%
           Ironwolf              44     2      4.6%
           Ironwolf Pro           8     1     12.5%
           Skyhawk               13     0      0.0%
           Exos X12               8     3     37.5%
           Recertified            1     0      0.0%
WD         Black                 13     0      0.0%
           Gold                  16     0      0.0%
           Purple                 4     0      0.0%
           Yellow (AE)            4     0      0.0%
           Blue (5400)           23     4     17.4%
           Red (5400)            38     2      5.3%
HGST       Deskstar              59     5      8.5%
HP         MB600X                 2     1     50.0%
DELL       M605ACA                4     4      100%
Marshall   Nireline               6     1     16.7%
TOTAL                           363    27      7.4%
 
 

Drive failures

 

The Dell and HP drives were quite a surprise. We couldn't find any detail on the Dell branded drives but for whatever reason, they were horrible. Yes, every single one failed.  No, I would never use one for production.  Also notice that there were no fails with the Toshiba drives! Nicely done, Toshiba!

Do keep in mind that any failures with a small number of drives (i.e. 50 or less) can easily show a disproportionately higher failure rate than should be expected in real world conditions. You can dive as deep as you want to in the statistical calculations but realize that a drive failure percentage for the Eros X12, for example, is inflated with such a small sample. There's no excuses for the Dell drives.

 

Making it happen

Drives waiting to receive their data.

Obviously, Defcon only runs for a defined time. To that end, we had less than 66 hours to accept, tag, copy, and return everyone's drives (1600 Thursday to 10 on Sunday). In addition, we were missing some duplicators this year leaving us with only 84 total drive copy slots).  Each 6TB copy cycle takes 15-20 hours to complete. Given those constraints, we buckled down to start duplicating.

We tried to get some additional duplicators at the last minute but failed.  With our capabilities known, we identifed our max duplication for the 66 hour period would only be about 308 drives (66 hours / 18 hours cycles * 84 bays). 

As we already accepted 327 drives in the first hour (an average of over 5 drives per minute), we stopped accepting drives early on Friday morning for a total intake of 473 drives.

We kept an extremely close eye on all of the duplicators for any drives slowing the process down <cough> Marshall </cough> and only left the duplicators with about 2-5 minutes of rest between cycles as we swapped drives out. By keeping things closely monitored 24x4, we were able to duplicate 319 drives this year. The duping broke down to look something like this:

  • Infocon.org archive:    99
  • Hash Tables 1 of 2:     121
  • Hash Tables 2 of 2:       99
  • Grand Total:                 319

Cranking away at duplicating Hash Tables

One side note: Too many people left their drives behind this year - about 30 were left over after the 11:00 pickup deadline!  We stayed as long as we can and were able to get all but about 10 of them back to their owners.  The remainder will be useful for the DDV next year!

 

Wrap up

In the end we were able to push about 1.9 Petabytes out the door to our community! A great result given the circumstances. Naturally, we want to do better next year.  We expect to have more duplicators and be able to take all comers in 2019 (which will be in Paris, Bally's, or Planet Hollywood next year).

If you're one of those who wasn't able to get your drive duplicated, all the content is now available through torrents or via Infocon.org. If you have the bandwidth, please seed them too!

Big thanks to DT, Miss Friday, Zant, the QM Stores, Goons, Jok, Jerry, Carlos, Bill, Jeff, and others for helping make DDV happen this year! See you in 2019!