Maximum reliability for your wordpress hosting

At wpdone we’ve gone to extreme lengths to engineer maximum reliability.

We distribute multiple copies of your website, not just to 1 server, but to 3. And not just to 1 data centre, but 3.

Our main data centre runs on industry leading VMware and a SAN. The other 2 datacentre’s are run by amazon. All servers and datacentre’s are in Australia.

We use a duplicated everything, shared nothing strategy. So no server or component is reliant on another server – they can function independently.

We use a load balancer between each layer to manage the availability of each component of each server.

Each component is also smart, and knows how to connect to each load balancer.

• we use technology from facebook to failover and replicate the memory cache between servers, and between data centres. Any web server can connect to any memory cache.
• we use a clustered database across the 3 datacentres. And web server can connect to any working database server.
• The security managers, and page optimization servers are also behind the balancers.
• We use smart routing.
  • incoming web requests can route around a failed component, using working components on other servers.
  • each internal request knows how to reach multiple load balancers.
    • we use a technology from wordpress.com hosting to enable wordpress to connect to all 6 database servers.
    • each wordpress instance can connect to all of the memory cache instances, in a failover and replication pattern
    • each security and page optimization server knows how to connect to each wordpress instance
    • each component knows to favour local resources if they are available, but during component outage or excess load they can access components on other servers in other data centres

By Scott Farrell on January 13th, 2016 , Follow @scott_WordPress Tweet to @scott_WordPress