Major failures of cloud services in 2012, and what conclusions can be drawn from this
Based on a recent report of the International Working Group on Cloud Computing Resiliency (IWGCR), cloud computing services are unavailable each year, on average, for 7.5 hours. Companies that partially or fully use the cloud for their applications and services have suffered several times this year. Let's look at the biggest failures in cloud services in 2012.
Microsoft Windows Azure
The largest and most extensive failure of Microsoft Windows Azure was in February, it affected all geographic locations, the full recovery of the service performance lasted more than 24 hours. Microsoft stated that the failure was caused by an error in the software associated with an incorrect calculation of the time and date for a leap year. The problem caused an angry reaction from the users of the service, who expected more coverage of the problem and more communication from Microsoft.
In July, Microsoft Azure cloud computing service was unavailable again, this time in western Europe, the failure time was 2.5 hours. The reason was in an incorrectly configured network device, which caused problems with connecting users.
Later in the fall, another failure occurred due to the work of Office365, during which millions of user mailboxes were unavailable.
Amazon Web Services
The power failure of Amazon Web Services in June cut off users from necessary services for 6 hours. The following services suffered: Amazon Elastic Compute Cloud, Amazon Relational Database Service and AWS Elastic Beanstalk located in the US East region of Virginia. In addition, cloud management companies and PaaS providers such as Stratalux, Digitaria, Heroku and PaaS service provided by Salesforce.com have suffered. Popular sites: Netflix, Pinterest, Reddit, Forsquare and Instagram also suffered from this crash.
Less than a month later, the second failure of AWS occurred, after which one of the major customers publicly announced that he was no longer using Amazon and was forced to look for alternatives.
Apple iCloud
In September, a large number of users of this cloud service could not access their mailboxes. The problem was related to the central iCloud service, so that it was common for users of Mac OS-based computers, iOS device users, and for users using the iCloud.com Web interface.
')
Google gmail
This year, the Google Gmail service suffered more crashes than in the past. The first failure occurred in April and lasted one hour. The problem touched less than 10 percent of users, the main reason was the incorrect configuration when performing a routine update operation of the system. The second failure occurred in June and hurt less than 1.5 percent of users.
Conclusion
Despite all the precautions taken by cloud service providers, disruptions occur regularly for various reasons, such as human errors, technical failures, or natural disasters. But this is no reason to abandon the clouds. All of these factors can be controlled using a comprehensive disaster recovery plan and the same resiliency plan. Over time, the reliability of cloud platforms will increase, and resiliency will strive for 100% (real, not announced by the marketing department). Sooner or later, iron hosting will be crowded out of the clouds.
Original article: www.rickscloud.com/major-cloud-outages-of-2012-to-learn-from
Posted by: Rick Blaisdell
Microsoft Windows Azure
The largest and most extensive failure of Microsoft Windows Azure was in February, it affected all geographic locations, the full recovery of the service performance lasted more than 24 hours. Microsoft stated that the failure was caused by an error in the software associated with an incorrect calculation of the time and date for a leap year. The problem caused an angry reaction from the users of the service, who expected more coverage of the problem and more communication from Microsoft.
In July, Microsoft Azure cloud computing service was unavailable again, this time in western Europe, the failure time was 2.5 hours. The reason was in an incorrectly configured network device, which caused problems with connecting users.
Later in the fall, another failure occurred due to the work of Office365, during which millions of user mailboxes were unavailable.
Amazon Web Services
The power failure of Amazon Web Services in June cut off users from necessary services for 6 hours. The following services suffered: Amazon Elastic Compute Cloud, Amazon Relational Database Service and AWS Elastic Beanstalk located in the US East region of Virginia. In addition, cloud management companies and PaaS providers such as Stratalux, Digitaria, Heroku and PaaS service provided by Salesforce.com have suffered. Popular sites: Netflix, Pinterest, Reddit, Forsquare and Instagram also suffered from this crash.
Less than a month later, the second failure of AWS occurred, after which one of the major customers publicly announced that he was no longer using Amazon and was forced to look for alternatives.
Apple iCloud
In September, a large number of users of this cloud service could not access their mailboxes. The problem was related to the central iCloud service, so that it was common for users of Mac OS-based computers, iOS device users, and for users using the iCloud.com Web interface.
')
Google gmail
This year, the Google Gmail service suffered more crashes than in the past. The first failure occurred in April and lasted one hour. The problem touched less than 10 percent of users, the main reason was the incorrect configuration when performing a routine update operation of the system. The second failure occurred in June and hurt less than 1.5 percent of users.
Conclusion
Despite all the precautions taken by cloud service providers, disruptions occur regularly for various reasons, such as human errors, technical failures, or natural disasters. But this is no reason to abandon the clouds. All of these factors can be controlled using a comprehensive disaster recovery plan and the same resiliency plan. Over time, the reliability of cloud platforms will increase, and resiliency will strive for 100% (real, not announced by the marketing department). Sooner or later, iron hosting will be crowded out of the clouds.
Original article: www.rickscloud.com/major-cloud-outages-of-2012-to-learn-from
Posted by: Rick Blaisdell
Source: https://habr.com/ru/post/163659/
All Articles