Its been a rough few weeks for the cloud. Amazon Web Services, the public cloud division of internet juggernaut Amazon, suffered a mighty blow, crippling the internet on a Tuesday afternoon, while internet-connected teddy bears were hacked.
If you could make God bleed, people would cease to believe in Him, Ivan Vanko tells Tony Stark in Iron Man 2. There will be blood in the water, the sharks will come. A simple typo brought the Internet to its knees on Feb. 28.
With a market share of roughly a third of the internet, cloud traffic relies on the AWS, particularly the Amazon Simple Storage Service. A tiny coding error brought down all of S3s customers, including popular services and websites such as Spotify, Imgur, Slack, Quora, and, ironically, the Down Detector.
This is the first major malfunction of S3 and the AWS, which is designed to deliver 99.999999999% durability for cloud services. Being one of the first cloud services commercially offered, Amazon quickly signed major names to the service. But, other providers, such as Microsoft, Google, and IBM, are catching up. Does this major outage equate to blood in the water, signaling a shift to other services?
Its possible, though unlikely. Given the rarity of AWS suffering a catastrophic failure, and the 10 percent reimbursement should the service drop below 99.9 percent availability, AWS will continue to be popular. For reference, its estimated the outage cost companies about $310 million. Thats not to say that companies will not decide to seek other options, as salespeople for other services have been given perfect fodder for pulling customers away.
Take, for instance, a small company rethinking their cloud service provider. If they were using Amazon, but only have a localized target base, they may decide that server geolocation is more important, speeding up load times. Amazons AWS is primarily based in northern Virginia, an area known as US-EAST-1, and the source of Februarys outage. This would not serve a Californian base altogether well.
The AWS was not alone in its cloud woes. Internet of Things-connected stuffed animals, CloudPets, from Spiral Toys, were hacked, leaking 800,000 user accounts. The database used to store CloudPets accounts was not behind a firewall, nor was it password-protected. Though the accounts themselves were hash encrypted, some were weak enough to be cracked. The data has been exposed since January, and the database was overwritten twice by hackers.
Worse yet, it was found the stuffed animals themselves could be remotely hacked, and turned into spy devices, or used to harass toddlers with audio output.
Despite this incident, cloud computing is very secure – often using the same encryption protocol as banks. Utilizing a third-party, offsite server can also make data recovery easier. Companies specializing in cloud data recovery can make what would otherwise be a catastrophic situation – such as a hack or onsite server malfunction – into an easily remediable situation. San Francisco Municipal, for example, used a cloud-stored backup to overwrite locked files when they suffered a system-crippling hack. Instead of paying a hacker a large ransom, they were up and running by the end of the weekend.
Though there is no acknowledgment on CloudPets site, and Spiral Toys stock is sitting has gone back and forth between a tenth of a cent and zero since February, answers to exactly what happened and why may not be forthcoming. Amazon, at least, detailed exactly what happened in a blog post.
The cloud may be the future, but problems will still arise. Look carefully over code for typos or errors, and always make sure your database and products are encrypted.