Could Yesterday's NHS Email Meltdown Have Generated As Much Data as the Large Hadron Collider Does In Over Two Months?

By James O Malley on at

Yesterday was a bad day for Britain's National Health Service. The NHS's entire email system was crashed when an IT Contractor in Croydon accidentally sent a test email to all 1.2 million members of staff.

To make matters worse, when people started to receive this rogue email they did something that has annoyed large organisations since time immemorial (well, probably the mid-90s): They hit reply all.

According to self-described "NHS Data Geek" Graham Hyde, as it currently stands there have been 394 reply alls. This means that there have been somewhere in the region of 474 million completely pointless emails sent around the NHS network since yesterday. No wonder the network crashed under the pressure.

What's astonishing though is when you consider just how much data transfer that involves. And according to some back of the envelope calculations with some slightly spotty assumptions, we reckon that NHS emailers have in the last day generated more data than the Large Hadron Collider does in over two months. That's right - uncovering the fundamental building blocks of the universe could actually be less data intensive than typing "testing 123" and hitting send before checking the recipient box.

So what are the numbers. Here's how we worked it out - which huge thanks to Richy Thompson and Andrew West, who did the sums first and spotted this surprising comparison.

1) If we make a conservative assumption that each email address contains 20 characters - that's 20 bytes each. (In actual fact, NHS addresses are probably longer because they contain the name of the section of the NHS they are in, and are formatted firstname.lastname@). This also means that this calculation doesn't include anything else about each email sent - nothing else in the headers, or the body, or the subject and so on.

2) Multiply that by 1.2m NHS staff that would have been in the headers of each email, that means each individual email was 24m bytes large - 24 megabytes each.

3) Multiply that again by 1.2m - as each 24mb email was sent to each of the 1.2m staff. That means that approximately 474 million emails have been sent in total.

4) That means that in total 1.1376E+16 bytes - or 11.376 petabytes (PB). That's 11,376,000 gigabytes. Or approximately one NHS network administrator's heart attack.

Compare that to the LHC, and according to the Worldwide LHC Computing Grid, scientists at CERN are expecting to generate 60PB of data this year. That means that assuming that the LHC generates data at a constant rate - these dodgy emails generated as much data as almost ten weeks worth of data collection on the Higgs Boson and the like. And it has managed this all in one day.

As I say, this is only back of the envelope stuff and the number in reality could vary wildly. The NHS uses Microsoft Exchange and Microsoft Outlook - wouldn't the latter crash if 1.2m email addresses were put in the send box? And could the email servers be doing something clever to compress or send bulk emails more efficiently? But either way, it still seems pretty clear why the entire system was taking down yesterday.

So what's the moral of this story? Check what you're sending before you hit send.