March 8, 2017,Posted by: Admin


Suspect EVM security means there will be doubts about all election results

If there is a single discrepancy found in any EVM count, one has to have an independent high-level inquiry and get to the bottom of it, find the cause and fix it. In the absence of it, there will always be doubts about any election results using these EVMs.

In 2010, a Hyderabad based security researcher Mr. Hari Prasad procured an Electronic Voting Machine (EVM) from somewhere ( probably picked up stealthily from the EVM storage with the help of someone) and showed two ways of hacking the EVM. He was imprisoned before the Supreme Court freed him.

He came up with a piece of hardware that is small enough to tuck in a shirt pocket and showed that with the help of that hardware and a brief physical access to the EVM machine, the votes can be altered.

In the recent Maharashtra civic elections, there have been reports of widespread discrepancies between the counts by the EVM and the total number of votes polled.

Another picture follows.

In both the cases shown above ( both pictures are picked from twitter and may need more authentication in a serious investigation, however now there are enough cases from highly credible sources to conclude that a prima facie case exists to conclude that EVMs have been tampered with ), the counts by the EVMs have been more than the total number of votes polled. Another case here this time from Uttar Pradesh buttresses the same hypothesis. However, this one could be due to a malfunction or is it preparation for a more sophisticated hack? Prima facie, in the first two instances it appears that more votes have been added into the EVM after the polling. Is there an inquiry? Not that I am aware of, as yet. Are there doubts raised about security of EVMs? None by the Election Commission, as yet, and here we have been using the same EVMs for election of 5 states, Uttar Pradesh India's biggest state included.

In another glaring case in the Maharashtra elections, a candidate got zero votes when he says that he voted for himself. See this link . He also says that his family would have voted for him. How can he then get zero votes? Has the election be countermanded ? Not yet.

In a fifth case, the dates on the EVM for some votes is 23 rd February while the date of polling is Feb 21, a clear indication prima facie that votes have been added to the EVM two days after the elections. Indian Express reports it here .

The point is that the design of the EVMs hasnt followed basic security practices that are prevalent worldwide. For instance, there is no encryption at all. The whole security depends on physical security of EVMs and the fact that no one should tamper with them from the time of voting to the time of counting. Yes, whether they work correctly at the time of voting is another issue. There is a mock trial done just before voting. That is the only assurance we have that the machine works correctly. However, mock trials wont show sophisticated tampering such as if ten votes are added after every 100 votes. This can be achieved by changing the firmware at the time of production. And there is no check whatsoever that the firmware in a machine is not tampered with at the time of production.

After the voting to the day of the counting, the EVMs are supposed to be stored in a strong room. But is there a guarantee that no one has access to the EVMs and this includes the Election commission personnel? In an incident in Punjab reported here some people went to the strong room in some pretext and started moving the EVMs.If one looks at the clarification, the officials state that the workers were moving EVMs related to some other election. This means that the building is not sanitized. Is the strong room sanitized? I know that candidates dont get a chance to check the strong room minutely. It is important that the strong room be thoroughly sanitized not just for presence of other EVMs but also for any backdoors including entry from the roof. Also, the candidates need to be allowed to put a tent just outside the strong room not just outside the building.

Finally, a seal on the EVMs with the signatures of various candidates or their election agents seems to be one of the protections. But then, signatures can always be forged. Another protection is the serial number on the EVM. Again, there is no guarantee that another EVM cannot be made with the same serial number.

Access to EVMs also has been a problem. A Varanasi poll official took a replacement EVM home and his son took pictures with it and posted on FB. See the link here . Thus, it is not impossible to have access to an EVM and get time to understand the circuitry and may be even reverse engineer the software of the EVM.

It has been also reported that some EVMs malfunctioned. An EVM was found in Assam which only voted for BJP. The link is here. As also, a similar case was found in Maharashtra. Did the EC get to the bottom of the problem and fix it? If they did, we dont know.

All in all, there are too many loopholes to trust the system. From a security point of view, it is primitive to put it bluntly.

Today, there is technology available to make things public as well as secure.

I will give you an example. Many softwares are kept for download on websites; many a times their sources ( that is, code) are also kept. One issue of concern is that some hackers change the softwares so that anyone who downloads software gets a hacked version. To counter that, there is something called a hash ( MD5 hash is a common type of hash used) which is some function of the bits in the software, so that it gives a unique result for the unique combination of bits in the software. Hash is just a 32 or a 64 bit number but due to the algorithm used to compute the hash, it has some unique properties. Another property of the hash is that for that hash, no other combination of bits will give the same hash when that algorithm is applied. Thus, if some hacker changes the software, the hash will differ and it will be known that the software is altered.

Thus, for instance, if the EVM had a hash calculating algorithm inside it which determined the hash of the votes, then if there was any alteration done to the votes from the time of polling to the time of counting, the hash would have differed and it would be clear that the votes have been altered. The hash could have been even made public and no one would know the votes as the algorithm is nearly impossible to reverse to get to know the votes from the hash. There would have been no need to be paranoid about security. Of course, the hash algorithm could be altered by hackers, but there are other ways to protect it. For instance, there are ways to check if the firmware in the machine is genuine or is altered. Again, encryption technology comes of use here. the hash of the code itself could have been taken, and any changes to the algorithm and hence to the firmware and to the code would have got revealed as it would produce a different hash.

The point is we are spending millions and billions of rupees on substandard technology and further we are not being transparent and open when genuine security researchers raise questions.

The Maharashtra EVM hack was crude. What if some hacker does a smooth job of it so that discrepancies are not noticeable?

All in all, if there is a single descrepancy found in any EVM count, one has to have an independent high level inquiry and one has to get to the bottom of it, find the cause of it and fix it. In the absence of it, there will always be doubts about any election results using these EVMs.

Election results are close. The only hope now seems to be the VVPAT which is the voter verifiable print audit trace. A VVPAT is a printout of the vote that the voter gets to see at the time of voting to confim that the button that he pressed indeed resulted in a vote to the candidate mentioned against the button. But if we are going to have to depend on it, that is an admission that EVMs have failed. Further, whether to allow counting of VVPATs is a decision that is at the descretion of a returning officer. And of course security of VVPATS is another topic altogether.

Some countries such as Netherlands and Germany have gone back to paper ballots. The question is : are paper ballots secure? There is no 100% guarantee unless there is a constant watch on them from the time of voting to counting. However, the long time lapse in current situation ( in Goa the elections were held on February 4) gives huge time for someone with an access to EVMs to do mischief. This is enough time to do even a research study and it is unlikely any party could guard the EVMs day and night non-stop for such a long time without slacking somewhere. In fact, this is time enough to even manipulate VVPATs. Thus, there is also an urgent need to keep the time period between elections and counting as little as possible.

PS: The Election Commission was not contacted for this article.


Copyrights © 2017 Teknotrends Software Pvt Ltd All rights reserved | Template by W3layouts