It began with a few dropped internet connections. Before long, some websites were persistently unavailable. For eBay sellers, things reached a crisis point when they found themselves unable to flog existing items or list new ones, and they started voicing their displeasure on social media.
The next day, breathless headlines informed us that Tuesday's outages were a consequence of the internet being "full up". It had, apparently, burst its digital zip. "Well, that seems rather serious," we thought to ourselves, trying to picture what on earth that meant in practice but failing to do so.
Our reliance on the internet, whether direct or indirect, is almost total, while our ignorance of the way it works is similarly all-encompassing. But that's hardly surprising. The chasm between the simple interfaces that we use daily and the sophisticated back-end technology that powers them is growing ever greater. We unquestioningly click on a friendly blue "log in" button, with no knowledge of, or indeed interest in, the complex procedures that deliver a "welcome" screen a couple of seconds later. We operate entirely on a "don't need to know" basis. In truth, we have little choice in the matter, but that suits us fine. Until something goes wrong.
Some attempts in the media to explain what happened on Tuesday were akin to the much-mocked remarks by former US Senator Ted Stevens, when he described the internet as a "series of tubes".
While my own understanding is compromised by my total lack of network administration experience, I'll give it a go. Forget tubes; the internet is more like a network of networks, all interconnected. Two tangentially related systems help us reach the website or internet service we want: one is the Domain Name System (DNS) that tells the packets of information where to go; the other is the Border Gateway Protocol (BGP) that tells them how to get there – which route to take across the internet. The problem was caused by the sheer size of the routing table, or map, created by the BGP; it's now so large that older routers used by big companies and internet service providers are becoming unable to store it; 524,288 table entries is the limit for an old router and that's what was "full up". Their response was either to crash, or slow down, or ignore new information; eBay was one of many websites to suffer as a consequence.
As it had nothing to do with our own devices or routers, it's not our problem to fix and in that sense we can breathe easy. Tuesday's slowdown was a consequence of big business failing to upgrade hardware or software to meet the challenges of a fast-growing internet. The problem was not unforeseen; router manufacturers such as Cisco have been warning businesses for months, but some were caught on the hop. All we care about, of course, is the likelihood of our digital lives being compromised again in the future – particularly when so much of it happens in a "cloud" over which we have little control.
Having someone to blame is always useful. When our "internet is slow" we call our ISP and complain, even if the problem has nothing to do with them. The aforementioned eBay traders furiously demanded compensation from the company on Tuesday without having any idea as to whether it was eBay's fault or not. But while we silently rage at the IT professionals to whom we seem beholden, the people who understand how this stuff works and seem to possess untold power, we should remember that they, in turn, are beholden to the people with the cheque books.
Stories have emerged of network admins pleading with their bosses to take notice of the BGP issue and invest in creaking infrastructure, only to be met by the same uncomprehending stare that we might give on being told that the internet is "full up". Thank goodness someone knows how it all works.