Google’s Gmail service is suffering an outage that has left 120,000 of its 150 million Gmail users without e-mail, contacts, labels and other content since the weekend. While some will no doubt use the incident as a basis to proclaim the cloud as unreliable, the truth of the matter is more complex. Both on-premise and cloud-based services can, potentially, suffer from data loss. With on-premise services, you (or your IT staff) are typically 100% responsible for backing up your data, securely storing it, and testing recovery procedures. All (or at least most) cloud-based services will take care of this for you; as part of outsourcing your IT to your cloud provider, you’re also outsourcing data backup, storage and recovery.
However, this doesn’t preclude the possibility of performing your own backups of cloud-based data. This is an important practice not only to protect yourself against data loss caused by the provider, but also against accidental data loss caused by you or your staff (with most cloud-based services the latter is much more likely than the former). Additionally, the proliferation of cloud-based data synced to mobile devices introduces the possibility of data data loss can also potentially occur if a mobile device is lost or stolen. To illustrate: a friend of mine recently had his iPad stolen. The iPad’s calendar was synched to Google Apps, and the thieves decided to wipe out the calendar to restore the iPad to a more sellable “blank slate” state. The Google Apps sync then deleted all of my friend’s calendar entries on Google, leaving him not only without an iPad, but without his calendar of his upcoming events. Because he hadn’t been backing up his Google Calendar, there was no way to recover the calendar.
Most cloud-based providers provide some mechanism to export data in an industry-standard format. For example, Contacts should be exportable in a CSV format, Calendars should be exportable in an iCal format. In the case of Google Mail or other web-based e-mail, use a desktop-based e-mail client to download your mails to your desktop.
Couple the backups of your cloud-based data with an additional backup to an external hard drive or, ideally, a cloud-based backup service, such as Mozy or Carbonite, and ensure you’re keeping at least 30 days of data backed up at all times.
With this two-pronged backup approach, if data is deleted from your cloud-based service – either by negligence or by accident – you have at least 30 days of backup snapshots to recover from. My friend with the stolen iPad could have simply restored his calendar from the day before the theft with this backup methodology.
Whether you are hosting data on-premise or in the cloud, data backup should be your number one priority. Make sure you’re regularly backing both your on-premise and cloud-based data on a daily basis, have copies of the data in multiple locations, and have data snapshots going back at least one week and, preferably, at least one month.