Slashdot Mirror
Sequoia Voting Systems Source Code Released
Mokurai sends a heads-up about Sequoia Voting Systems, which seems to have inadvertently released the SQL code for its voting databases. The existence of such code appears to violate Federal voting law: "Sequoia blew it on a public records response. ... They appear... to have just vandalized the data as valid databases by stripping the MS-SQL header data off, assuming that would stop us cold. They were wrong. The Linux 'strings' command was able to peel it apart. Nedit was able to digest 800-MB text files. What was revealed was thousands of lines of MS-SQL source code that appears to control or at least influence the logical flow of the election, in violation of a bunch of clauses in the FEC voting system rulebook banning interpreted code, machine modified code and mandating hash checks of voting system code." The code is all available for study or download, "the first time the innards of a US voting system can be downloaded and discussed publicly with no NDAs or court-ordered secrecy," notes Jim March of the Election Defense Alliance. Dig in and analyze.
Appears to control or at least influence the logical flow of the election
What exactly does that mean?
Means they suspect that the code for the actually tallying and evaluating ballots is in SQL. It is suggested that this violates the law for being dynamic and interpreted.
The file they have is simply a SQL Server backup.
It takes a few minutes to restore using SQL 2005 Express + SSMSE
Nothing has been destroyed or sabotaged.
but...
When the database is restored you get the tables with the data in. :)
All the stored procedures have been deleted. Or so Seqoia thought
As the use of strings on the backup file demonstrates, the text of the sp's are still there.
There are various database tools (Lumigent was one from memory) that allow looking back through the database log and, I expect, returning the database to a previous state.
Just when companies had got the hang of cleaning up after track changes they move on to SQL database backups :)
Maybe it's a cultural thing, but I've never seen the necessity to complicate things any further than paper, pencil, double physical count. Cheap, no machines involved, fast. On a national election down here (about 15 million voters), voting booths close at 6pm and results are known nation wide right on time to open the 8pm evening news.
Except that Americans like to vote on everything.
Not just politicians, but sherifs, judges, district attorneys (i.e., head government prosecutors), etc. Add this to the fact that most elections (municipal, county, state, federal) tend to happen on one day, so that when you walk into the booth, you don't just have a piece of paper, but a small booklet to go through. Then add propositions (i.e., referendums) that many states have if enough people sign a petition. If you want to be an educated voter on all the possible choices you have to do some serious studying.
And then you have to count all of these 20+ separate run offs for the various levels of government.
How to restore the .bak file using Microsoft SQL Server Express 2008:
Step 1. Go download SQL Server Express 2008 (This is trivial, left up to the reader. You might have to go to a microsoft webpage) and install.
Step 2. Go download SQL Tools for SQL Server (Trivial) and install.
Step 3. Go download the .bak.zip file from the above wiki. Save it to 'C:\foofoo\'. Unzip the .bak file within it to 'C:\foofoo\'. You should now have: 'C:\foofoo\RIV_20081104_Canvass_Final_dbset_E.bak'
Step 4. Start up SQL Server Express
Step 5. Open SQL Management Studio and connect to your local SQLEXPRESS instance.
Step 6. Click on the top most node in (Should be your machine's name\SQLEXPRESS). Click "New Query".
Step 7. Run the following query:
RESTORE DATABASE RIV_20081104_E FROM disk='C:\foofoo\RIV_20081104_Canvass_Final_dbset_E.bak'
WITH MOVE 'RIV_20081104_Esys' TO 'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\RIV_20081104_Esys.mdf',
MOVE 'RIV_20081104_Edat' TO 'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\RIV_20081104_Edat.mdf',
MOVE 'RIV_20081104_Elog' TO 'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\RIV_20081104_Elog.ldf',
REPLACE
go
Step 8. Wait.
Step 9. This should create a database called RIV_20081104_E.
Have fun.
Show me the section in federal law that cites wiki!
OK, I'll quote this instead:
"4.2.2 Software Integrity
Self-modifying, dynamically loaded, or interpreted code is prohibited, except under the security provisions outlined in section 6.4.e [sic - see note below]. This prohibition is to ensure that the software tested and approved during the qualification process remains unchanged and retains its integrity. External modification of code during execution shall be prohibited. Where the development environment (programming language and development tools) includes the following features, the software shall provide controls to prevent accidental or deliberate attempts to replace executable code: ...
IANAL, but that seems pretty clear cut. Maybe not in regards to the SQL, but as far as to what they mean.
First, I'm the guy that built that wiki page.
Second, "code that defines races" can be used to alter results. I have a lot of experience playing with Diebold databases because we've had access to those since 2003 when Diebold left an FTP site open. If you swap the candidate ID numbers between two candidates in the Diebold database (run in MS-Access), you'll flip the election. In a heartbeat.
It *appears* there's code present in this Sequoia database to do the same thing. Note the word "appears". The best way to find out, and the most MORAL way, was to put it up for public review.
Risking exposure of our technical warts, sure. Still worth it. Check the discussion areas at the wiki - we're learning a hell of a lot, very quickly.
But yes, it's true: I don't know MS-SQL, and nobody else at EDA does either. So we were faced with a choice: find a few people who did know it, pay 'em a bunch of donated money to write a formal report behind closed doors, or do a public review and exam even if that means exposing any mistakes we make, knowing they'll be caught pretty damn quick.
Which was better?