Burn baby burn… So I am sure you all have been aware of the AWS S3 outage by now and various comments on twitter – #awsoutage
If you have a website, chances are you have been or were affected in some way shape or form. A part of Amazon Web Services infrastructure had issues (S3) which is pure essence in part or Public Cloud infrastructure that hosts pictures, videos, images and files/storage for a vast amount of services you currently use (Airb&b, SoundCloud and Slack to name a few).
[easy-tweet tweet=”We need to rethink the way we consume cloud services” user=”neilcattermull” hashtags=”Cloud, AWS”]
Public Cloud is a very difficult model to provide as a service when you consider the reasons why to take it in the first place.
Public Cloud is a commodity for technical infrastructure – requiring an inexpensive service to gain your business – that translates to an inexpensive solution to resiliency to keep the costs down.
Public Cloud needs to be flexible – requiring a multitude of variables of storage, types of files, bespoke code and to have the ability to grow quickly – very hard to cater for absolutely every connotation of provisioning safely, securely at a price that’s competitive.
Being competitive on price always comes with a cost operationally and this recent outage at AWS has demonstrated this. Event the webpage that is provided by AWS to monitor the systems and locations of the service was hosted on the same infrastructure that went out of service. There have been a few very large IT providers that have simply pulled out of providing Public Cloud infrastructure provisioning as they see this as fools gold. To provide a resilient, secure, multi-tenanted and flexible service is a very hard task to do in the best of examples but by adding in the low-cost element too simply makes it a business model that just isn’t viable (commercially).
We need to rethink the way we consume cloud services to not just be about “how much it costs” to “what happens if the service goes down? (and it will)”.