Earlier this month, Microsoft 365 users were unable to log in to multiple Microsoft 365 services that use Azure Active Directory, including Outlook, Microsoft Teams, Office.com, Power Platform and Dynamics 365. Users already logged in were unaffected and the services themselves continued to run but for most or all of the five-hour window anyone logged out was unable to log in.
Microsoft is not the only one. Adobe Creative Cloud also went down early this month. At the end of September, several Google services went offline or were barely reachable, including Google Meet, Drive, Docs, Analytics, Classroom and Calendar. In early June, IBM reported a multi-hour, multi-region cloud services disruption.
In fact, a quick search shows that cloud service outages are not uncommon. The problem is that for enterprises that depend on these services, outages are not just time consuming but also have a significant impact on the bottom line. So, should you trust them?
Cloud Outages Are Rare
Recent high-profile cloud outages have raised concerns about over-reliance on cloud services, said Ted Rohm, ERP analyst at Canada-based Technology Evaluation Center. Though worrisome, serious outages are relatively rare and the benefits from using cloud services outweighs the risk.
“The fact that they occur doesn’t negate the wide-ranging benefits of cloud-based business software and should not deter organizations from pursuing cloud offerings," he said. "It should instead be a reminder that even the largest and most experienced global software providers may experience downtime.”
Instead, Rohm said outages should remind companies to:
- Seek solutions built on stable cloud infrastructures to minimize such outages.
- Have emergency procedures to maintain critical business operations during outages.
Related Article: The Role of Distributed Cloud Computing in the Enterprise
Build a Backup Plan for Cloud Outages
Pieter Vanlperen, founder and managing partner at PWV Consultants, a New York City-based tech trouble-shooting consultancy, pointed out that even though the big three cloud providers have dealt with service disruptions recently, cloud providers can still be trusted to provide what they said they will provide.
An outage is not a common occurrence and, even though it is frustrating for businesses, it should not cause a loss of trust in the provider. Enterprises should have a backup plan for any vital or mission-critical business functions.
"Nothing that is essential to the functionality of your business should be built atop a cloud service without having a backup in place,” Vanlperen said. "With proper backups, business can continue in the event of a disruption. It really does not matter which provider is being used, outages can happen to any company at any time."
If the outages continue to be rare, there is no reason to lose trust in them because outages can happen to any company. He added that no matter which provider you choose, it is important to have proper backups in place.
Related Article: Is It Time to Revise Your Business Continuity Plan?
No Choice: Cloud Services Essential to Remote Work
As many employees continue to work remotely, companies really do not have a choice. They must rely on cloud-based tools. The takeaway is that enterprises should not rely on a single set of applications, especially when everything is from the same provider or developer, said Morten Brøgger, CEO of San Francisco-based Wire.
Learning Opportunities
“Over-reliance on a single suite not only allows for massive disruption of workflow productivity in the event of an outage, it can also open companies up to serious security risks. It is crucial for companies, especially in this era of remote work, to have a backup plan and a method for employees to safely communicate in the event of a service outage,” he said.
If they do not, communications will likely move to less secure or unsanctioned channels, thereby opening the risk of possible cyberattack or loss of company intellectual property or data. Imagine how devastating it would be for a company to experience a collaboration suite outage, followed immediately by a data breach.
As an aside, Brøgger said that on-premise communication software is a good tool for companies to have in their stacks because it can be installed locally in company computers and servers rather than being based in the cloud.
The Hybrid Solution: On-Premise and Cloud Data Centers
So, is there another way? Asaf Ezra, CEO and co-founder of Israel-based Granulate, suggested a solution that is becoming increasingly popular with enterprises, especially those working in highly regulated industries.
He pointed out that cloud environments provide flexibility and the ability to quickly scale as traffic surges. He also noted that using your own data center is approximately four times less expensive than the cloud, so organizations must ask themselves when and if it is the right time to consider a hybrid strategy to benefit from both flexibility and cost optimization.
To benefit from both, organizations that already have on-premise data centers might consider keeping those and using the cloud only for excess capacity. In addition, organizations that can might consider setting up at least some of their environment as on-premise, ideally to handle average traffic, and expand to the cloud when needed to handle peak traffic.
“The rising popularity of hybrid environments is especially interesting to watch from the cloud providers’ point of view," Ezra said. "The big three are all selling managed on-premise solutions (AWS with AWS Outposts, Microsoft Azure with Azure Arc and GCP with Google Anthos) and constantly investing in these.”
They understand that if customers are hooked into their API for managed on-premise, those customers will use the same provider if they choose to expand. “So, it is safe to say that whoever controls the data centers controls the market, from migration, cost and future standpoints. You lock in your customers by selling on-premise,” he said.
One appeal of the cloud was as a multi-regional option. The need to avoid long response time, which impacts performance and hurts the quality of service, might be impacted by 5G and reduced network latency. The round trip from Australia to the U.S. will still be too long, but lower overall latency will effect the need for multi-regional presence. Time will tell what impact this will have on the need for multi-regionality.