Could the limits of cloud hosting be reached during the pandemic?

By | April 19, 2020

Last week, Microsoft’s collaborative cloud service, Teams, was unable to withstand the onslaught of requests from millions of users who were called upon to telework en masse around the world with the COVID-19 health crisis. “2020 will be a terrible year for IT services,” warned the IT research and analysis company, GlobalData. However, while declines in demand are to be expected, the need for cloud resources is expected to continue to explode, driven not only by the current crisis and teleworking constrained for many, but also the acceleration of digital transformation projects leading inevitably more and more companies hitherto reluctant to throw themselves into the vast ocean of the multicloud, to go there.

“As companies stop their brick and mortar operations and, where possible, move to a remote workforce, it’s clear how important the cloud is to business continuity. Any organization that actively resists digitization now faces a harsh reality. This puts cloud providers in a strong position,” said Chris Drake, senior analyst, Global IT Technology and Software at GlobalData. The consequence for cloud providers is far from trivial, as it raises the question of their ability to support the rising burden of remote access.

Creating Impacted Azure Virtual Machines

Could the situation become critical? There are individual signals to be taken into consideration: in Europe, for example, many Azure customers have reported capacity limits for opening new virtual machines or even starting them up. However, until now, sizeable incumbent cloud providers offering computing, storage, and networking services have been better able to handle the increased usage. “Leading public cloud providers, such as Microsoft Azure, AWS, and Google Cloud, have built huge global networks that are incredibly well equipped to handle traffic spikes. So, not surprisingly, we have seen virtually no COVID-19-related impact within their networks. Considering the number of outages detected over the last six weeks, the trend line remains relatively stable globally, with only a slight increase in the United States,” ThousandEyes says.

It remains to monitor the evolution of this trend over time closely, knowing that for other IT and telecom service operators, the situation appears more worrying. “Failure incidents within large UCaaS provider networks are relatively rare; however, the recent massive increase in usage highlights the current design limitations. Capacity would be increased across the board to meet new service demands,” ThousandEyes warned in a note. Over the past 24 hours, 31 outages in 23 locations have been reported worldwide, affecting a wide variety of players ranging from Akamai to Cogent to AT&T. In this context, in France in particular, organizations such as Arcep and the French Telecommunications Federation have taken the lead in urging major video service providers – led by Netflix and Google YouTube – to make efforts to limit their bandwidth consumption. A call that was heard even in the ears of Walt Disney, who agreed to postpone the launch of its Disney+ streaming service by 15 days in France, but not in 7 other European countries such as Great Britain, Germany, and Italy, despite being cruelly affected by the coronavirus SARS-Cov-2.