Serverless Backups Outside of AWS

Simon Bennett
Simon Bennett · Feb 16, 2022
Serverless Backups Outside of AWS Artwork

Back in June last year we launched our compute platform to enable you to backup resources without requiring your own server, think things like Github Backups or managed MySQL where you just don't have a server.

Or like in SnapShoots case all our servers are behind firewalls and temporary, over the course of a day we can scale from a handful of servers to dozens and there is never a guarantee a server will be there again to connect to, while you could have your own dedicated server for backups, with SnapShooter, Compute you was not required, you could use our service.

We span up an AWS Fargate container to perform the backup and you just paid for the CPU/RAM and networking inbound if connected to an AWS resource.

But there was a massive catch, you had to store your backups in AWS S3, via your own or via SnapShooter Storage. This was a huge limitation which at the time we thought we could live with, but on reflection, we are not happy with it.

So we are very proud to announce we have built an alternative, SnapShooter Serverless, our previous project with Fargate will remain and be renamed SnapShooter AWS compute, and will still have its place for AWS based workloads. For people wanting to store their backups outside of AWS S3 or have workloads that originate outside of AWS our new compute option will be for you.

How does SnapShooter Serverless work?

From your perspective, it's totally serverless, but from ours, we are managing a fleet of dedicated servers that provide the horsepower to perform the backup. We have partnered up with DataPacket a UK based hosting company to provide our compute, we choose them due to their amazing network performance, each one of our servers has a 10gbit uplink. You can tell they have good networking due to being the company behind CDN77.

We have decided to roll all costs into one metric, networking, You pay a fixed cost based on the amount of data you egress from our serverless network, prices are currently set to $0.03 per GB, this includes the timeshare of resources to pull, compress and upload your data. If you are processing more than 3TB of data a month you can talk to sales about data commitments.

Over the coming weeks we will talk more about what it took to build our fleet, we have started in Frankfurt, due to its amazing networking interconnect in Europe/UK, but will be expanding to the US in the summer. We will take about our decision to purchase IPv4 space as well as support IPv6 with our new network and where we see the future of serverless going.