Hollywood is fond of recycling the apocalypse in all its flavors. In almost 90 years of filmmaking, Earth has been torched, flooded, battered by asteroids, attacked by aliens, ravaged by disease, decimated by conflict, hit by solar flares, overrun by AI and zombified (the most implausible apocalypse but a firm Hollywood favorite).
It is still missing Outage, a buttock-clenching tale of the chaos that unfolds when an all-powerful corporation accidentally brings down the Internet. But real life has already advanced the narrative. On December 7, Amazon blamed "network device issues" for one of the worst Internet blackouts ever.
Dozens of firms that rely on the AWS public cloud were affected, from Disney to Tinder. Between 9.37 AM and 4.35 PM Pacific Standard Time, games stopped working, TV screens went dark, online sales ground to a halt and Internet-enabled household appliances refused to cooperate. The only thing Hollywood needed that was missing were the subsequent scenes of looting and violence, of desperate netizens fighting on the streets for survival.
What appeared to be a seven-hour glitch – roughly the same duration as Facebook's a few weeks ago – was never going to end civilization. But it was the rudest reminder yet of the planet's dependency on technologies that hardly anyone understands, controlled by a small number of trillion-dollar firms.
A smaller company might have seen investors look for sanctuary elsewhere. Not Amazon, whose share price even gained 3% on outage day. Frankly, there are few other places to go. Synergy Research reckons AWS controlled about one third of the global market for cloud infrastructure services in the recent third quarter. With shares of 20% and 10% respectively, Microsoft Azure and Google Cloud make up the trio that serves 63% of all business. A major outage at any of those companies could be economically devastating.
For an indication of just how fast the world's transition to the cloud has occurred, refer to Amazon's results. Back in 2013, when Barack Obama sat in the White House, the AWS unit was an interesting growth story generating about $3.1 billion in annual sales. Last year, it made nearly $45.4 billion.
Anyone that has plunged into the public cloud should be worried not just about market dominance but also the difficulty of switching providers. No one has spelt this out more clearly than Snap, the company behind the Snapchat messaging service and a tenant of both AWS and Google.
Noting that its systems are "not fully redundant" on the two cloud platforms, Snap went on to say in its last annual filing with the US Securities and Exchange Commission that: "Any transition of the cloud services currently provided by either Google Cloud or AWS to the other platform or to another cloud provider would be difficult to implement and would cause us to incur significant time and expense."
Between 2017 and 2022, Snap has made commitments to spend about $3.1 billion with Amazon and Google, a figure that – when annualized – equals about a fifth of Snap's entire revenues for the 2020 fiscal year. "Given this, any significant disruption of [sic] or interference with our use of Google Cloud or AWS would negatively impact our operations and our business would be seriously harmed," said the company.
None of that has dissuaded Dish Network, the company building America's fourth mobile network, from entrusting just about all its IT systems and network functions to the AWS cloud. AT&T, perhaps more worryingly (as a company that already serves millions of customers), has decided Microsoft will look after its 5G core.
The backlash begins
But a European telco backlash against these sorts of deals has begun. "Hope the folks at AWS fix their big problem and re-light their big candle – in the meantime I re-refer you to this," tweeted Neil McRae, chief architect of the UK's BT. His other tweet, the one to which he was re-referring his followers, reads: "So still want to put your network core into the public cloud? #suckers."
Hope the folks at AWS fix their big problem and re-light their big candle - in the meantime I re-refer your to this: https://t.co/zA3zDPJkbf— Neil J. McRae (@neilmcrae) December 8, 2021
Scott Petty, the chief digital officer of Vodafone, has been similarly scathing. "Our view would be that's too risky and that you are almost outsourcing a core competency," he said at a recent press event when discussing the AT&T arrangement with Microsoft. "You need to be able to work effectively with all the key players and move workloads around."
So determined is Vodafone to avoid Snap's fate that it has even started investing in its own software tools, allowing it to move IT workloads from one environment to another. Over the next few years, it plans to add another 7,000 software engineers to the 9,000 it currently employs.
European stakeholders, meanwhile, are pushing ahead with Gaia-X, a vague plan to create sovereign data infrastructure for Europe. It has been championed by bigwigs such as Timotheus Höttges, the CEO of German telco incumbent Deutsche Telekom. "If we find partners here in Europe who are driving this Gaia-X or the open-source standard for cloud infrastructure, this might help us as well in the edge environment," he told analysts on a recent call.
Any initiative that might produce sustainable alternatives to AWS, Microsoft and Google – or make it easier to shift from one to another – probably deserves encouragement. At a time when it's fashionable to play at being Nostradamus, one reasonably safe prediction for next year is that an even worse outage lies ahead.
- AWS outage cuts down swaths of the Internet
- Facebook outage should frighten everyone
- Akamai outage is latest warning about Internet dependency
- Vodafone plans software blitz to cut IT spend and cloud risks
- Prickly Deutsche Telekom boss wants hyperscaler deals curbed
— Iain Morris, International Editor, Light Reading