Apparently the root cause of the outage that stopped Exchange Online for North American users last week has been traced to a failed Cisco router or other component. Sources say “networking gear” but aren’t more precise and Microsoft isn’t saying. A failure of such a basic infrastructure link that doesn’t seem to have been backed up with some redundancy is embarrassing all round and not what customers expect from high-quality datacenters designed to deliver essential services such as email.
Microsoft says that the datacenter’s networking facilities have been remediated (fixed in layman’s language) and that they continue to monitor the network very closely. Of course, you’d expect them to have been monitoring the network very closely before the outage occurred to identify component failures but maybe now they’re looking at each individual packet as they arrive or leave the datacenter. We shall see in due course.
I think that Microsoft management has done the right thing by moving immediately to compensate customers for the effect of the outage.A note emailed to customers said:
“We understand that any disruption in service may result in a disruption to your business. As a gesture of our commitment to ensuring the highest quality service experience Microsoft is proactively providing your organization a credit equal to 25% of your monthly invoice. The credit will appear on a future invoice, and you do not need to contact Microsoft to receive this credit.”
Microsoft’s own data indicates that the outage lasted 190 minutes (11:30AM to 2:40PM PDT) so there’s no way that they can meet their “financially backed, guaranteed” SLA of 99.9% availability for the month of August. However, handing over a credit of 25% is a pretty big chunk of revenue for Microsoft to give back to customers. I don’t know how many customers have signed up for Office 365 nor do I know the split between North American users who were affected by the outage and those, like myself, from the rest of the world who continued to work throughout. It’s therefore impossible to calculate a precise figure for the revenue loss. However, you can speculate in an attempt to determine a ball park figure, so here goes.
Let’s imagine that Microsoft has signed up 1 million paying users for Office 365 since its formal launch and that the average price paid per user is $10/month, equivalent to the price of the entry-level plan for enterprises (E1).We further assume that the user population is split evenly between North America and the rest of the world. The monthly revenue for North America is therefore 500,000 x $10 = $5 million. Giving back 25% of that will cost Microsoft $1.25 million. the revenue lost could easily be higher as Office 365 plans are priced at up to $27/month and of course more users might have been affected and so qualify for the rebate. In any case, you could buy a lot of network redundancy or even a few network administrators for the kind of money that Microsoft has had to pay to maintain customer confidence.
Joking apart, I like the fact that Microsoft moved to lance the boil that might have otherwise festered by paying out so quickly. This will reassure customers and demonstrate Microsoft’s seriousness about running Office 365 as a high-quality utility service. The only pity is that the outage had to happen to demonstrate just how much a problem can cost.