How an expired credit card can shut down your entire business

Written by David Mytton

Back in the old days when you had physical servers hosting your website, a hosting provider would have to take physical steps to suspend a customer’s account – blocking network access or powering down equipment. In the virtualised/cloud era, it’s now a case of issuing a single command to halt or delete your instances.

Instead of paying manually by invoice through an accounts receivables team where due dates and multiple attempts to collect payment would be made, billing is now completely automated through credit card. You could forget to pay off your card limit. Your card could be stolen. It might expire. You could be on holiday and not read the payment alerts. Or something could simply go wrong with your bank.

Amazon has disabled my account because of a billing screwup by AMEX. Where should I buy my ebooks while they sort their crap out?

— Will Critchlow (@willcritchlow) February 7, 2013

This means there’s potential for a failed payment to become a single point of failure as your hosting provider suspends your account for overdue invoices. Do you know how many times your vendor will attempt to bill your account? Will they just suspend your instances or will they be deleted? Is this the same across your hosting vendor? E-mail? DNS?

X (software/infrastructure/platform) as a service has many benefits but have you considered billing as a single point of failure?

Enjoy this post? You may also like Making a point with SLAs

Alex Kinch

Funnily enough I was reading a similar thread about Softlayer the other day whilst doing some research after reading about your decision to leave Rackspace for Terremark and later to SL. As a Rackspace customer a scenario like this concerns me; after reading this thread: http://www.webhostingtalk.com/showthread.php?t=1230289 about Softlayer I struck them off my list. Dedicated self owned infrastructure is where we need to be.

Ps: enjoy the blog, have read it end to end. Love sysadmin Sundays.

https://blog.serverdensity.com David Mytton

For us it may be different because we spend so much with them and have a dedicated account manager in Europe and the US, and multiple points of contact on our account. I suspect it was a combination of that guy having just a single server, expired card and incorrect contact details on file. Even if they didn’t try and call him, they do send automated e-mails for billing notifications so they must’ve been missed or not delivered.

The size of the customer account shouldn’t make a difference to the initial procedure (e-mails, attempts to call) but if they have exhausted all forms of contact there’s not much else. Physical hardware has a cost (it’s not the same as just suspending a software service).

I’d suspect the same thing would’ve happened with any provider.