MM007 – Deploying low-cost Disaster Recovery on Oracle Cloud

Pubblicato da Saverio Leoni il

When approaching Disaster Recovery solutions most of companies tend to view it as a cost, but when things go wrong, what are you going to do? This is the reason you need to carefully plan a way to quickly regain operativity.

But at what cost?

If you hire physical devices in a data center you would need hardware performance comparable to the one you’re using in production, or when you’ll go disaster nothing will work. Very little companies are able to (or can) compartmentalize the processes and choose the essential – and run only the essential – when a disaster comes. This causes a waste of resources when you don’t need them.

Fortunately Cloud solutions are able to overcome these issues. You can rent VMs with small shapes when on standby operativity (and the I/O performance on archive apply is more precious than CPU) and then change shape when you’ll need it.

We successfully applied this with one of our customers, using Oracle Cloud and cutting monthly costs to one third of the original on-premises solution.

What about performance?

Unfortunately for the customer we had a chance to test the DR solution as soon as the core part of it was implemented. They had a very serious problem -caused by a bug- at the primary DB site which lead us to a switchover of all the DBs with little to none advance notice.

Result: even with no application servers on cloud (just the DBs) the company managed to work for three days without interruption, while we fixed the primary site.