What? Wait…

I was reading about the Iowa Caucus.

The acting DHS secretary Mr. Wolf said that the application issue appeared to be a “Load” issue. By “Load” he means that the servers were unable to keep up with the number of requests.

Okay I’ll buy that is a possibility, if everyone in a state was voting at the same time. But as the number of voters decreased, the server would catch up and post each transaction in turn. If this was the problem then it’s pretty obvious whoever tested the software didn’t do any load testing and quite possibly didn’t do much testing at all.

Lets face it, we’re all familiar with online opinion polls, and I’d imagine the servers handling those are dealing with millions of votes a minute. Seems to me that Shadow (The company that apparently spent 3 years building the software,) would have looked to other examples of voting systems, during their development process.

For god’s sake, there are PORN sites that handle votes for performers without crashing. Given the prevalence of Porn Sites I’d guess they process something on the order of MILLIONS of votes per Second.

Then I read that the application was only for the 170 – 190 precinct captains. So the paper votes were cast, then counted, and the captains were to use an application to input those numbers?

You’re telling me that with 3 YEARS of development no-one ever tested with a measly 200 simultaneous users?

WTF?

As I sit reading more about this, I’m astounded.

I have Apple Time Capsules here in my home that can handle 50 simultaneous users on WiFi.

A low end Dell server purchased from Best Buy could probably handle 250 users from the moment it was plugged in, possibly more if all the server had to do was tally incoming data for ONE Single application.

I have to point out that I’m kinda talking out of my hat here because I don’t have all the facts. So take what I’m saying here with a salt lick.

My point is, that with something as important as votes, if I could put a system together with commercial of the shelf (COTS) equipment for less than 10K in hardware and a little web programming there is absolutely NO EXCUSE for the debacle we saw in Iowa.

Much less so when you factor 3 YEARS of development time.

Hell, with 3 years of development time, I could give you Web and Phone based access, Live updates, and auditing of figures entered by precinct, candidate, and user. Complete with state of the art security. I’d have also taken the DHS up on testing the system too. The DHS has an entire division dedicated to Cybersecurity. 

I’d probably have requested that the FBI and NSA take a look too, if they were willing.

WHY?

Because the product would have to be rock fucking solid and more eyes looking at a system are more likely to find flaws that can be corrected before its debut.

Especially given that over the last four years we’ve heard about nothing but Russian influence in our election process. I’d be wanting to make something that was so secure that there’d never be any question about the veracity of the product or its results.

Make no mistake, this is (or was) a product.

Shadow would have been in a prime position to resell the product to all 50 states and would have been reaping the benefits for decades with maintenance and upgrade contracts.

Now Shadow will fade into the morning light like a bad dream, having made millions (I’m guessing) for its principals and casting everyone below executive level to the unemployment line.

Oh, and they’ll have an added lovely parting gift of FAILED project on their resumes.

As I said, we don’t yet have all the facts and likely, we never will.

Online voting could be a reality. But only if we commit to doing it right. 

Don’t you find it interesting that we have more security in place online and over the phone to deal with our banking needs than we do when dealing with the direction of our country as a whole?