Recovering data after a 100 year flood

[Preface: Unitrends has a quite a few customers who have a disaster but are reticent about publicly disclosing this information.  And yet new buyers quite often ask us for this kind of information.  For that reason I’m going to share stories about Unitrends customer disaster recovery events that have occurred; but I’m going to do so in a manner in which specific customers can’t be identified.  Note that all customers involved have approved of this “anonymous” sharing of information.]

Unitransport has been a Unitrends customer for more than 5 years.   Uptime is critical to the operation of Unitransport; without their production systems they can’t keep their logistics systems operating for more than a few hours.  Unitransport is a multi-site company with IT professionals at two of their sites.  Initially Unitransport used Unitrends with multi-site replication to provide disaster recovery.  Because the IT leader and professionals that work for her are deeply paranoid (in a good way) about data loss, they actually used both multi-site replication (hot backup copies) and cold backup copies (archives via rotational disk.)  The reason was that they wanted to ensure that in the case of a disaster that they had working backup in the worst case.

Unitransport was one of the first customers of Unitrends Cloud.  Executive management had made the decision to save costs by moving to a single centralized IT team. The reason they adopted Unitrends Cloud was that they wanted to lower their IT operational expense associated with second site multi-site replication and rotational archiving.  Unitrends Cloud allowed them to do this.

Unitransport’s environment consisted of 20TB of data of which 18TB was virtualized using VMware.  The other 2TB consisted of physical Windows servers.  There were 20 virtual machines; the latest virtual machine was 3TB.  There were 3 Windows physical systems.

Unitransport had a designated for disaster recovery in Unitrends cloud an 8-hour RTO (Recovery Time Objective) SLA requirement on most of these systems but had a 1-hour SLA requirement on a selected set of these systems. .  The company had a 24-hour RPO (Recovery Point Objective) for Unitrends Cloud and had a 2-hour RPO for local continuity

There was a “100-year storm” that flooded their data center.  Unitransport called Unitrends and declared the disaster; within one hour all of their systems were operating and their VPN (Virtual Private Network) connections were made.  The company’s physical data center had been switched to a virtual data center residing within Unitrends Cloud.  Operations in Unitrends Cloud continued for 10 days; the company selected one of their other locations as their temporary data center and Unitrends shipped physical appliances from our cloud to Unitransport’s temporary data center in a “reverse seed.”  Operations resumed at this temporary data center for Unitransport.

Unitransport eventually moved IT operations back to their primary data center.  They credited Unitrends with both keeping their business going as well as making the move to the temporary data center (and eventually their primary data center) relatively stress free.  The term “relatively” is important – in talking to the leader of IT she admitted that this was one of the most stressful events of her professional career.

We love comments. Leave a Reply.