[CAVO] [VVSG-election] [VVSG-interoperability] Single Point of Failure - the Scan Head - RE: By November, Russian hackers could target voting machines

Thu Jul 28 17:35:38 UTC 2016

Great point Susan. This is why we print a paper ballot. Specifically, we print a ballot on a single, blank sheet. The printed ballot is the ballot of record. So, if we are using COTS, even if it's 100% foreign, it doesn't store a ballot locally on the machine and it's not connected to the network. If someone hacks the code, they can't change the outcome of the election such that it's undetectable. This is the equivalent of me taking your ink pen, breaking it and asking you to mark the ballot without noticing your ink pen doesn't work properly.

Thanks,

On 7/28/16 1:15 PM, Brent Turner wrote:
Susan--

What is your alternative to open source code ?   Closed / disclosed ?  The DOD / NASA / Air Force etc side with open source ...

I think the question regarding best approach open vs closed has been called and is now long over--

Regarding audit capability.. the printed ballots themselves are the countable record..

Best-  BT

On Thu, Jul 28, 2016 at 8:31 AM, Susan Eustis <susan at wintergreenresearch.com<mailto:susan at wintergreenresearch.com>> wrote:
 I agree this is attractive, and should be done, but what about a usable audit trail?  GPL v3 open source with COTS hardware has enormous potential for buried hack handles in the code and the COTS has BIOS vulnerability.  I contend with that in my business all the time, the coders in foreign countries leave hooks in the open source code that is not detectable and that may not be activated for years.
Susan

On Thu, Jul 28, 2016 at 11:20 AM, Susan Eustis <susan at wintergreenresearch.com<mailto:susan at wintergreenresearch.com>> wrote:
There is plenty of precedent in the court cases in Massachusetts where that did happen.

Susan

On Thu, Jul 28, 2016 at 10:43 AM, Arthur Keller <ark at soe.ucsc.edu<mailto:ark at soe.ucsc.edu>> wrote:
Thanks, John.

Some people would worry that the delay in reporting would allow nefarious activities to occur.

Best regards,
Arthur

On Jul 28, 2016, at 7:38 AM, Wack, John (Fed) <john.wack at nist.gov<mailto:john.wack at nist.gov>> wrote:

I hesitate to jump in, but I do agree that it’s very very important to get the election night count accurate.  I tend to think that no matter what, it’s a human process, a logistical nightmare for some locales, and that this needs to be recognized in technical discussions and recommendations.  One of the best things to do, in my opinion, to make things more secure and more accurate, falls into the usability category for tabulation: perhaps election results shall not be released until such and such a time on the next day, say noon.  Pressure to get everything correct in a very short amount of time after a long day works against security and accuracy.  No matter what technology we use or how secure or whatever, it can’t erase the fact that the overall process is a very demanding management issue.

The CDF work can help to make that easier, common identifiers for geopolitical geography and contests can help to make it easier, and there are probably a host of other items that could help to reduce the amount of time (and software required) to get all the equipment to work together smoothly.  So as I’m reading the posts, I’m thinking about future VVSG requirements to make all the equipment work together so that the overall tabulation process is more usable to the election people conducting it.  Of course, giving them more time to do would help significantly.

Cheers, John

From: vvsg-interoperability-bounces at nist.gov<mailto:vvsg-interoperability-bounces at nist.gov> [mailto:vvsg-interoperability-bounces at nist.gov] On Behalf Of Arthur Keller
Sent: Thursday, July 28, 2016 10:20 AM
To: Stephen Berger <stephen.berger at suddenlink.net<mailto:stephen.berger at suddenlink.net>>
Cc: vvsg-election <vvsg-election at nist.gov<mailto:vvsg-election at nist.gov>>; vvsg-pre-election <vvsg-pre-election at nist.gov<mailto:vvsg-pre-election at nist.gov>>; vvsg-post-election <vvsg-post-election at nist.gov<mailto:vvsg-post-election at nist.gov>>; vvsg-interoperability <vvsg-interoperability at nist.gov<mailto:vvsg-interoperability at nist.gov>>
Subject: Re: [VVSG-interoperability] Single Point of Failure - the Scan Head - RE: By November, Russian hackers could target voting machines

Thanks, Stephen. I think you mean scanner software is NEVER examined.

Best regards,
Arthur

On Jul 28, 2016, at 7:02 AM, Stephen Berger <stephen.berger at suddenlink.net<mailto:stephen.berger at suddenlink.net>> wrote:
Susan,

Good points.

Let add to the landscape that currently we have a single point of failure that I think deserves some attention.  That point is when the ballot is scanned.  Typically the scan mechanism is not specified and the initial processing software is not specified.  Almost all, actually to my knowledge all of the scanners first throw away a lot of information.  Modern scanner electronics is able to get excellent resolution and color differentiation.  There is a lot that can be done with high quality can images.  However, the scanner software immediately throws away most of that information and make everything black or white.  This is done because it makes mark recognition easier and it saves on machine memory.  However, what is a vote is determined not off the image of the ballot but the processed image.  To make matters worse, neither the VVSG or any election official, decides when a pixel should be determined to be black or white or how many pixels make a valid mark.  This is left to each company and even each design team at each company.  Even worse, it is often decided by the scanner engine manufacturer and that software is very examined in any of our processes.

It would seem worth paying some attention to what happens between the ballot being feed to the scanner and a decision being made about what votes are on that paper.  It also seems reasonable that election officials should be the ones deciding how big a mark is a valid mark and how various kinds of uncertain marks should be dealt with.

Best Regards,

Stephen Berger
TEM Consulting, LP
Web Site - www.temconsulting.com<http://www.temconsulting.com>
E-MAIL - stephen.berger at ieee.org<mailto:stephen.berger at ieee.org>
Phone - (512) 864-3365<tel:%28512%29%20864-3365>
Mobile - (512) 466-0833<tel:%28512%29%20466-0833>
FAX - (512) 869-8709<tel:%28512%29%20869-8709>

From: vvsg-interoperability-bounces at nist.gov<mailto:vvsg-interoperability-bounces at nist.gov> [mailto:vvsg-interoperability-bounces at nist.gov] On Behalf Of Susan Eustis
Sent: Thursday, July 28, 2016 8:31 AM
To: Arthur Keller <ark at soe.ucsc.edu<mailto:ark at soe.ucsc.edu>>
Cc: vvsg-election <vvsg-election at nist.gov<mailto:vvsg-election at nist.gov>>; vvsg-pre-election <vvsg-pre-election at nist.gov<mailto:vvsg-pre-election at nist.gov>>; vvsg-post-election <vvsg-post-election at nist.gov<mailto:vvsg-post-election at nist.gov>>; vvsg-interoperability <vvsg-interoperability at nist.gov<mailto:vvsg-interoperability at nist.gov>>
Subject: Re: [VVSG-interoperability] By November, Russian hackers could target voting machines

Arthur, I agree, I concur.   My new book lays this scenario out in detail and provides suggestions for preventing the hacks, ways to protect the integrity of the election results, there needs to be safe guards and automatic recounts the very next day with observers representing all candidates, no matter whether the election was close or not.  There needs to be an audit trail and a way to protect the integrity of the balloting that occurs before election day.  There needs to be a way for the observers to make a duplicate of the original ballots as the recount goes on and to run those through their own counting scanner to determine the validity of the election.  There needs to be a way to interrupt the recount at any time if someone has to go to the bathroom or falls asleep so that the recount process has continuity and integrity.  Things like this.
Susan

On Thu, Jul 28, 2016 at 9:22 AM, Arthur Keller <ark at soe.ucsc.edu<mailto:ark at soe.ucsc.edu>> wrote:
But vote tabulation and especially roll up is often connected to the Internet. And with the lack of effective audits in more jurisdictions, hacking the Internet-connected vote tabulation systems would do the trick.

In particular, if the vote tabulation system is connected to the web reporting system, then that's an avenue for attack.

There's a difference between auditable and actually audited. If the results are sufficiently skewed on election night, post election audits may not matter anyway. They didn't even matter in Florida in 2000 where the election was close.

Could the programming of electronic voting machines be hacked in a Stuxnet type attack while they are loaded with the election data file?

If China can hack Google, do we really believe there's no way Russia can't hack enough counties or states to change the outcome of the presidential election?

Best regards,
Arthur

On Jul 28, 2016, at 6:07 AM, Deutsch, Herb <hdeutsch at essvote.com<mailto:hdeutsch at essvote.com>> wrote:
Voting machines are not attached to the internet.  You can’t hack them without physical control and that is auditable.

From: vvsg-interoperability-bounces at nist.gov<mailto:vvsg-interoperability-bounces at nist.gov> [mailto:vvsg-interoperability-bounces at nist.gov] On Behalf Of Arthur Keller
Sent: Thursday, July 28, 2016 12:30 AM
To: John Wack
Cc: vvsg-election; vvsg-pre-election; vvsg-post-election; vvsg-interoperability
Subject: [VVSG-interoperability] By November, Russian hackers could target voting machines

What should the election community do about this threat?

Best regards,
Arthur

https://www.washingtonpost.com/posteverything/wp/2016/07/27/by-november-russian-hackers-could-target-voting-machines/

By November, Russian hackers could target voting machines
If Russia really is responsible, there's no reason political interference would end with the DNC emails.
<image001.jpg>
By Bruce Schneier July 27 at 3:10 PM
Bruce Schneier<https://www.schneier.com> is a security technologist and a lecturer at the Kennedy School of Government at Harvard University. His latest book is Data and Goliath: The Hidden Battles to Collect Your Data and Control Your World<https://www.schneier.com/book-dg.html>.
Russia was behind the hacks into the Democratic National Committee’s computer network that led to the release of thousands of internal emails just before the party’s convention began, U.S. intelligence agencies have reportedly<http://www.nytimes.com/2016/07/27/world/europe/russia-dnc-hack-emails.html> concluded.
The FBI is investigating. WikiLeaks promises<http://www.cnn.com/2016/07/26/politics/julian-assange-dnc-email-leak-hack/> there is more data to come. The political nature<http://www.defenseone.com/technology/2016/07/how-putin-weaponized-wikileaks-influence-election-american-president/130163/> of this cyberattack means that Democrats and Republicans are trying to spin this as much as possible. Even so, we have to accept that someone is attacking our nation’s computer systems in an apparent attempt to influence a presidential election. This kind of cyberattack targets the very core of our democratic process. And it points to the possibility of an even worse problem in November — that our election systems and our voting machines could be vulnerable to a similar attack.
If the intelligence community has indeed ascertained that Russia is to blame, our government needs to decide what to do in response. This is difficult because the attacks are politically partisan, but it is<http://arstechnica.com/security/2016/06/guest-editorial-the-dnc-hack-and-dump-is-what-cyberwar-looks-like/> essential<https://www.balloon-juice.com/2016/07/26/we-are-at-cyber-war-so-what-exactly-do-we-do-about-it/>. If foreign governments learn that they can influence our elections with impunity, this opens the door for future manipulations<http://www.huffingtonpost.com/michael-gregg/top-six-ways-hackers-coul_b_7832730.html>, both document thefts and dumps like this one that we see and more subtle manipulations that we don’t see.
Retaliation is politically fraught and could have serious consequences, but this is an attack against our democracy. We need to confront Russian President Vladimir Putin in some way — politically, economically or in cyberspace — and make it clear that we will not tolerate this kind of interference by any government. Regardless of your political leanings this time, there’s no guarantee the next country that tries to manipulate our elections will share your preferred candidates.
Even more important, we need to secure our election systems before autumn. If Putin’s government has already used a cyberattack to attempt to help Trump win<http://talkingpointsmemo.com/edblog/trump-putin-yes-it-s-really-a-thing>, there’s no reason to believe he won’t do it again — especially now that Trump is inviting the “help.”<https://www.washingtonpost.com/politics/democratic-national-convention-obama-biden-kaine-set-to-tout-clinton-as-commander-in-chief/2016/07/27/afc57884-53e8-11e6-bbf5-957ad17b4385_story.html?hpid=hp_hp-top-table-main_trump-1230pm%3Ahomepage%2Fstory>
Over the years, more and more states have moved to electronic voting machines and have flirted with Internet voting. These systems are<http://arstechnica.com/tech-policy/2015/04/meet-the-e-voting-machine-so-easy-to-hack-it-will-take-your-breath-away/> insecure<https://www.statslife.org.uk/significance/politics/2288-how-trustworthy-are-electronic-voting-systems-in-the-us> and<https://www.salon.com/2011/09/27/votinghack/> vulnerable<https://www.theguardian.com/us-news/2015/apr/15/virginia-hacking-voting-machines-security> to<http://whowhatwhy.org/2015/08/31/foreigners-could-hack-us-elections-experts-say/> attack<http://www.popsci.com/gadgets/article/2012-11/how-i-hacked-electronic-voting-machine>.

[Your iPhone just got less secure. Blame the FBI.<https://www.washingtonpost.com/posteverything/wp/2016/03/29/your-iphone-just-got-a-lot-less-secure-and-the-fbi-is-to-blame/>]
But while computer security experts like me<https://www.schneier.com/blog/archives/2004/11/the_problem_wit.html> have sounded<https://www.giac.org/paper/gsec/3687/inherent-problems-electronic-voting-systems/105962> the<http://homepage.cs.uiowa.edu/%7Ejones/voting/congress.html> alarm<https://cs.stanford.edu/people/eroberts/cs181/projects/2006-07/electronic-voting/index_files/page0004.html> for<https://citp.princeton.edu/research/voting/> many years, states have largely ignored the threat, and the machine manufacturers have thrown up enough obfuscating babble that election officials are largely mollified.
We no longer<https://scontent.xx.fbcdn.net/hphotos-xlp1/v/t1.0-9/12115815_699872940152206_2266030088084252627_n.png?oh=2a4e5e944a5feadb7e133dd8c57be376&oe=57AD8C92> have time<https://xkcd.com/463/> for that. We must ignore the machine manufacturers’ spurious claims<https://www.salon.com/2006/09/13/diebold_3/> of security, create tiger teams to test the machines’ and systems’ resistance to attack, drastically increase their cyber-defenses and take them offline if we can’t guarantee their security online.
Longer term, we need to return to election systems that are secure from manipulation. This means voting machines with voter-verified paper audit trails<http://votingmachines.procon.org/view.answers.php?questionID=000291>, and no<http://engineering.jhu.edu/magazine/2016/06/internet-voting-nonstarter/> Internet<https://www.verifiedvoting.org/resources/internet-voting/vote-online/> voting<http://www.scientificamerican.com/article.cfm?id=2012-presidential-election-electronic-voting>. I know it’s slower and less convenient to stick to the old-fashioned way, but the security risks are simply too great.
There are other ways to attack our election system on the Internet besides hacking voting machines or changing vote tallies: deleting voter records<http://thehill.com/policy/cybersecurity/278231-election-fraud-feared-as-hackers-target-voter-records>, hijacking candidate or party websites, targeting and intimidating campaign workers or donors. There have already been multiple instances of political doxing<https://www.schneier.com/blog/archives/2015/11/the_rise_of_pol.html> — publishing personal information and documents about a person or organization — and we could easily see more of it in this election cycle. We need to take these risks much more seriously than before.
Government interference with foreign elections isn’t new, and in fact, that’s something the United States itself has repeatedly done<https://www.lawfareblog.com/what-old-and-new-and-scary-russias-probable-dnc-hack> in recent history. Using cyberattacks to influence elections is newer but has been done before, too — most notably in Latin America<http://www.bloomberg.com/features/2016-how-to-hack-an-election/>. Hacking of voting machines isn’t new, either. But what is new is a foreign government interfering with a U.S. national election on a large scale. Our democracy cannot tolerate it, and we as citizens cannot accept it.

[Why would Russia try to hack the U.S. election? Because it might work.<https://www.washingtonpost.com/posteverything/wp/2016/07/26/why-would-russia-interfere-in-the-u-s-election-because-it-usually-works/>]
Last April, the Obama administration issued<https://www.whitehouse.gov/blog/2015/04/01/our-latest-tool-combat-cyber-attacks-what-you-need-know> an<https://www.whitehouse.gov/blog/2015/04/01/expanding-our-ability-combat-cyber-threats> executive<https://medium.com/the-white-house/a-new-tool-against-cyber-threats-1a30c188bc4#.jgbalohyi> order<https://www.whitehouse.gov/the-press-office/2015/04/01/executive-order-blocking-property-certain-persons-engaging-significant-m> outlining how we as a nation respond to cyberattacks against our critical infrastructure. While our election technology was not explicitly mentioned, our political process is certainly critical. And while they’re a hodgepodge of separate state-run systems, together their security affects every one of us. After everyone has voted, it is essential that both sides believe the election was fair and the results accurate. Otherwise, the election has no legitimacy.
Election security is now a national security issue; federal officials need to take the lead, and they need to do it quickly.

--
--

Susan Eustis
President
WinterGreen Research
6 Raymond Street
Lexington, Massachusetts
phone 781 863 5078<tel:781%20863%205078>
cell     617 852 7876<tel:617%20852%207876>

--
--

Susan Eustis
President
WinterGreen Research
6 Raymond Street
Lexington, Massachusetts
phone 781 863 5078<tel:781%20863%205078>
cell     617 852 7876<tel:617%20852%207876>

--
--

Susan Eustis
President
WinterGreen Research
6 Raymond Street
Lexington, Massachusetts
phone 781 863 5078<tel:781%20863%205078>
cell     617 852 7876<tel:617%20852%207876>

--
Juan E. Gilbert, Ph.D.
Andrew Banks Family Preeminence Endowed Professor & Chair
Computer & Information Science & Engineering Department
Herbert Wertheim College of Engineering
University of Florida
P.O. Box 116120
Gainesville, FL 32611
352.562.0784 (V)
352.273.0738 (F)
juan at ufl.edu<mailto:juan at ufl.edu>
Twitter: @DrJuanGilbert
http://www.juangilbert.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/cavo_lists.opensource.org/attachments/20160728/557035d0/attachment.html>