Skyline Architecture

How is BC and DR arranged with Skyline

A simple, high-availability and quickly rebuild-able Load Balancer routes traffic to a Primary application/database server. A Backup server, on which the application is updated daily and the database is updated in real-time, runs at another data centre. The separate data centre is for the purpose of ensuring the two systems never end up on the same server host. If the Primary server goes offline, the Load Balancer can be pointed to the Backup server and the application is run from there. Once the Primary server is back online, this is brought up to date with code/data and becomes the Backup server, as this is the fastest method of having the full topology back online. It’s worth noting the Backup server is of the same specification as the Primary server, and so can run the application at full speed. If the Backup server goes offline, this can be brought up to speed when it comes back online, or rebuilt, whilst the Primary server continues unaffected. In the very unlikely event the Load Balancer goes offline, it is a basic machine that can rebooted in less than a minute, or it can be rebuilt from a snapshot in minutes, if required.

To protect against user-initiated data damage (a high level access user, against warnings/protections performs action that results in data loss), hourly database backups are taken on the Backup server and kept for 24 hours. Beyond 24 hours, a daily backup can be used.

System Resilience

All Digital Ocean data centres provide 99.99 SLA. The single point of failure is the Load Balancer itself, which can be rebuilt/rebooted very quickly, if necessary.

Data Centre Details / Where is it located?

London (Equinix) /Netherlands (TelecityGroup)

Are there back up locations and where are those?

Primary Server (London): 7 days – Nightly Backups

Backup Server (Netherlands): 24 hours – Hourly Backups / 7 days – Nightly Backups

Nightly Backup Server (Backupsy – London):30 days – Nightly Backups

Topology Diagram