Business Impact Of A Cloud Crash

It's a key question every business needs to ask, regardless of who its potential cloud service provider is - especially businesses that are more risk-averse.

If this question was posed to Microsoft or any knowledgeable Office 365 vendor, the most likely answer would be that "It won't." or "It's highly unlikely." Let's examine the reasons for these replies and then return to the key question.

VIPs with the distinct privilege of being allowed to visit a Microsoft Office 365 data centre, would be impressed by the sheer size and amount of neatly organised technology. However, they would probably not be able to see the vast amount of redundancy designed into the data centre to cater for practically any disaster, which includes acts of God as well as man-made. In short, Microsoft has redundancy implemented throughout all layers of the network to ensure its Office 365 cloud-based services are highly available and always on-line. Here are some examples of this redundancy.

  • Network interface card (NIC)-teaming for each server to ensure rapid failover should a network card go wrong.
  • Multiple top-of-rack (TOR) switches to provide redundancy for each server rack should a network switch fail.
  • Two devices used for routing and switching, and all connections are on a redundant basis.
  • Duplicate systems with automatic failover are configured for all firewall and load-balancer deployments.
  • Two separate network connections for each customer environment in the managed network.
  • Two individual power feeds for each managed network.
  • Redundant, high-capacity (n x 10GE) links into the Microsoft backbone for each data center. These links provide protected connectivity to the Internet edge and to other Microsoft locations.

There's more redundancy in network card, server, switch, firewalls, routers, power feed and incoming Internet feeds, the list goes on, but the ultimate redundancy exists by way of the geo-redundant data centres. In other words, in Europe, the Office 365 data centres between Amsterdam and Dublin will seamlessly failover to the other should one data centre experience a complete disaster. This failover is tested regularly.

So the key question now is, "what is the financial impact from Microsoft's perspective if, in the highly unlikely event, all data centres fail simultaneously?"

According to Kevin Allison, General Manager of Office 365 and Exchange, in his video on Business Impact:-

"We do actually provide service SLAs to our enterprise customers. This is a financially backed guarantee where we're saying we'll achieve the following service level and if we fail, we'll pay you back. That payback is actually in service credits or real dollars based on how they're engaged with the service."

The Microsoft Office 365 service level agreement (SLA) guarantees a 99.9 per cent uptime. That means, if the terms of the SLA are not met, you are eligible to receive service credits equal to a percentage off your total monthly bill. Should the monthly uptime be less than 95 per cent, then you can obtain a 100 per cent service credit for the month of use, per affected employee. As of writing, these are the service credit tiers for any loss of monthly uptime.

-- Monthly Uptime Percentage-- --Service Credit--
25%
50%
100%