Business Cloud News
Amazon suffered a partial or 'gray' outage

Amazon suffered a partial or ‘gray’ outage

Amazon on Sunday suffered a short outage in its web services operations that served as a reminder of the occasional perils of outsourcing services to the cloud.

An Amazon Web Services (AWS) datacentre in North Virginia suffered a hardware issue on a networking device which caused some hosted instances to become unavailable. The incident lasted just under an hour but also dragged down some notable hosted services such as Instagram and Vine with it.

The episode followed a similar event the previous week, which lasted for 25 minutes and had an effect on Amazon.com’s retail site in North America.

Many cloud service providers like Amazon often talk about the importance of geographic redundancy – advising customers to spread their services across multiple datacentres. The North Virginia AWS is the company’s biggest and oldest and has faced issues before. What’s clear though is that many companies, even big web stars, aren’t making their services redundant for whatever reason. Some commentators claim redundancy is still too complicated and others that it’s still too expensive.

What’s interesting though is that Instagram, owned by Facebook, and Vine, owned by Twitter, still rely on Amazon for service delivery. Twitter is no stranger to service issues despite (or because of) its popularity. But Facebook has extensive experience in building its own datacentres and fibre networks so it might be that this latest Amazon incident spurs on the moving of Instagram to in-house resources.

Comments
  • Debra Wilkins August 28, 2013 at 4:23 pm

    Hey James, Outages in a cloud or any networked environment can and do happen but I believe that the benefits far outweigh the potential downtime. But if companies are smart they will understand what an outage means to their users, customers, business etc. There are tools out in the market that will replicate a cloud network, including users, components, applications etc., understanding how downtime is experienced and what applications and users experience can really help manage a potential situation and also having this knowledge will really enable you to set-up a realistic SLA with your potential provider.

  • Post a comment

    Threaded commenting powered by interconnect/it code.

@BizCloud
News