With recent high profile outages, businesses are considering whether they should look into cloud outage insurance. But in my view, purchasing cloud provider insurance is only really putting a plaster over the wound – the fundamental issue to be addressed is whether your cloud provider is capable of providing the availability you need, even if there’s a disaster that affects them.
The best way to ensure that your provider can do this is to look at their track record. A few months ago, on the two hottest days of the year, there were significant outages in two major data centres in Docklands, which affected large parts of the internet, as well as several major providers. As a matter of fact, Maintel also use those very same data centres to provide services to customers.
However, our customers were not significantly affected, because we deliberately build redundant systems across multiple data centres in order make sure we carry on operating even when something like this happens. Providers need to ensure that they are prepared for the worst for their customers, instead of requiring customers to spend more money on insurance.
Another important aspect is the lead time on recovering: if your cloud provider suffers a major, sustained outage, and is not capable of restoring the service, then having a chunk of money to rebuild may be useful, but the lead times on replacing equipment is typically weeks, not hours – and that’s before we even think about the lead time on connectivity. And a DR plan based on an insurance payout can’t be tested.
A DR plan based on moving to a different cloud provider, or using a provider that builds in redundancy can be tested – and may also be cheaper than buying insurance.