Last June AT&T announced plans to shift its 5G core network to Microsoft Azure’s cloud, essentially putting the company’s core network in Microsoft’s hands. Likewise, Dish Network, which is in the midst of building a standalone 5G (5G SA) network, plans to host its 5G core and radio access network (RAN) in Amazon Web Services (AWS) cloud, which means it is entrusting its core network to a cloud provider as well.
AT&T’s and Dish Network’s alignment with the public cloud providers is part of a larger trend that is expected to increase in the future. Analysys Mason recently released its telecom predictions for 2022 and one of the research firm’s predictions is that more telecom providers will form multi-year partnerships with the cloud companies as they make critical decisions about deploying 5G SA and moving to a cloud-native platform.
“AT&T’s landmark 5G network deal with Azure will also push CSPs in the direction of public cloud providers in 2022 as they make their decisions about the cloud-native platform that they should deploy,” Analysys Mason said.
But AWS’ December 7 outage, which lasted more than five hours and impacted customers using certain application interfaces in AWS’ Eastern U.S. region, surely has to give wireless operators that are thinking about having public cloud providers host their core networks a bit of a pause.
The AWS outage knocked out many services including Venmo, Disney Plus and Tinder, and even put some Amazon deliveries on hold. And this wasn’t the first outage for AWS. The company experienced a similar multi-hour outage in November 2020 that also knocked out many sites and services.
Wireless network operators spend millions every year in advertising dollars touting their network reliability and coverage. Any network outage is a problem. But one that lasts several hours and happens with some frequency is not going to sit well with users, particularly those enterprises that have been guaranteed (and are paying for) a high level of network quality of service.
But are these cloud outages enough to make wireless operators reconsider their cloud strategies? Roy Chua, co-founder and principal at AvidThink, said he believes it is a concern for operators but they have to weigh the pros and cons. As Chua noted, wireless operators themselves aren’t free from experiencing network outages. In fact, T-Mobile was fined $19.5 million by the FCC for a massive 12-hour outage that occurred in June 2020 and led to more than 20,000 failed 911 emergency calls.
“Each will have outages,” Chua said. “It’s not clear to me that the cloud is risky or that it increases the risk of an outage.”
Chua also noted that the big cloud providers have been recruiting telecom network talent away from the wireless operators for the past few years and that means that over time the cloud providers may eventually have more expertise in keeping wireless networks up-and-running than wireless operators.
Of course, one of the other issues with AWS’ recent outage is that it took a few days before the company offered any explanation for what happened. AWS said in a post on December 10 that the outage was caused by a glitch in the company’s internal network that hosts “foundational services” such as application and service monitoring, the AWS internal Domain Name Service (DNS), authorization and parts of the Elastic Cloud 2 (EC2) network control plane.
Basically, there was “a surge in connection activity that overwhelmed the networking devices between the internal network and the main AWS network and that resulted in delays for communication between these networks,” AWS said, adding that it has already taken several actions to prevent this from happening again.
Chua added that the fallout from the AWS outage might have been alleviated if the cloud provider had been more transparent about the cause of the outage instead of waiting a few days to offer any explanation.
And he said that if the cloud providers want to secure multi-year contracts from wireless operators that want to move their 5G core networks to the cloud and move to a cloud-native platform, they are going to have to make sure that there is clear communication between the two companies and that transparency is not an issue.
It’s probably unrealistic to think that network outages will never happen. It is realistic for customers to assume that there is going to be quick recovery from any outage and that any network failure will be handled with transparency.