Widespread AWS Outage Silences Alexa; Solution Providers Critical Of AWS Response


A widespread Amazon Web Services outage Friday prompted a more-quiet-than-usual Alexa as Amazon's personal assistant technology, which is housed on AWS servers, appears to have succumbed to the regional outage.

Users began reporting problems with Amazon's Alexa and Echo speakers around 9:30 a.m. EST, according to website Down Detector. By Friday afternoon, Alexa and Echo outages were being reported throughout the U.S. with the East Coast and California appearing to take the biggest hits.

[Related: 8 AWS Offerings Gaining Popularity Right Now]

Amazon, for its part, acknowledged a networking error at 10:30 a.m. EST stemming from its multiple Virginia-based data centers. The company, however, did not comment or respond to CRN's request for comment on the Alexa outage.

Sponsored post

While Amazon does not have a status update page dedicated to its consumer products, business customers also rely on Amazon's Alexa technology. In November, Amazon unveiled Alexa for Business, a service designed to simplify common business tasks, such as setting up reminders, inventory management, and managing agendas and calendars.

Alexa for Business is available through select AWS solution providers.

Some reports have suggested that the Alexa outage was caused by an AWS service outage impacting the voice recognition and natural language processing servers that power Alexa.

Users also reported issues or downtime Friday with other third-party applications that live on AWS servers such as team collaboration application Slack.

Users are reporting that Alexa running on Echo speakers, as well as third-party devices, is giving replies such as: "I am not sure what went wrong" or "Sorry something went wrong" or a total loss of connection with no response whatsoever.

John Shaw, founder and CEO of Digital Nebula, an Austin, Texas, artificial intelligence startup that is pioneering an AI-based ELE.AI cloud ecosystem for solution providers, said the Alexa outage is a warning shot of sorts on the dangers of AI and natural language processing in an Internet of Things world.

If AWS had combined Alexa for Business with its $1 billion acquisition of smart doorbell maker Ring, an outage could spark a break-in at homes that are using the technology, said Shaw. "This is a huge problem," he said. "If a criminal knew that Alexa and Ring were down, they could break into your home."

If Alexa for Business was used in a self-driving car and the outage caused an accident, the question would be who is ultimately responsible: "AWS or the insurance company?" said Shaw.

The lesson to be learned from the outage, Shaw said, is the mantra often preached by AWS CTO Werner Vogels: Build for failure. "This shows AWS is not taking into account the fundamental message from its own CTO," Shaw said. "This may be a human error or something else; whatever it is they need to fix this and get a status page for Alexa and Echo. They need to build for failure."

Digital Nebula, for its part, is building an AI-based system for solution providers designed to work in a multicloud environment to avoid problems on a single platform like AWS. "What we are trying to do is make sure we build a system that is resistant to this type of failure," he said.

Raymond Tuchman, CEO of Experis Technology Group, a fast-growing Potomac, Md.-based HPE private cloud powerhouse with its 80,000-square-foot cloud services data center, said the AWS outage hammers home yet again the benefits of private cloud versus public cloud.

"This is more firepower for us to go tell the private cloud story with higher uptime, better security and better redundancy and at least a 40 percent cost savings compared with AWS," he said.

Tuchman said more customers are starting to listen to the benefits of private cloud versus public cloud, but it requires a lot of education. "A lot of customers go to the public cloud without doing the research to compare private cloud uptime, security and redundancy," he said.

One of the big advantages of Experis versus AWS, said Tuchman, is he and his company's responsiveness to customers. "You can call me 24 hours a day, seven days a week," he said. "AWS does not even have a status page for consumers using Alexa and does not pick up the phone. That is a problem."

The AWS outage comes just after onetime AWS stalwart Dropbox revealed in its S-1 Securities and Exchange Commission filing that it saved $74.6 million in operating costs over two years by moving from public cloud storage to its own lower-cost, custom-built infrastructure.

HPE was the driving force behind the Dropbox move two years ago to a hybrid cloud model with an "open architecture approach and a custom server solution" with HPE ProLiant and Cloudline service provider systems -- all financed by HPE Financial Services.

STEVEN BURKE contributed to this story.