This leading Alternative Investment Fund Management (AIFM) company has a strong policy how to deal with a potential large-scale disaster: should the headquarters be subject to an event where potentially the building (or the whole city) is eradicated, operations need to be able to resume within three days. Periodic review of SLAs and tests revealed that the existing disaster recovery site needed a hardware refresh, the lease was due to expire on the colocation. Plus on top the current tape backup solution was cumbersome and error prone.

English PDF Version

German PDF Version


During the initial meeting, comdivision’s Yves Sandfort and Philip Kriener learned from the infrastructure and end-user-computing managers, that the current disaster recovery solution consisted of 8 nodes plus dedicated legacy storage located in a colocation space at a local service provider, which was not easy to keep up to date. The compute infrastructure was connected to the on-premises datacenter with two dedicated lines. The replication of critical applications was done using Veeam.

The Challenge

The current backup and disaster recovery platform was both complex and expensive. A hardware refresh was due soon and if a disaster occurred, where the backup solution was to be scaled up to maintain operations, the system was difficult to scale up.  

Should a larger disaster happen, where all of the city’s infrastructure might be compromised, operations at remote offices were in jeopardy. That also applied to the Veeam backups, that where stored in a tape library at the colocation, not ideal for remote recovery.

After engaging with the customer in a workshop to establish a mutual understanding of the customer’s needs, Yves Sandfort, comdivision’s lead architect on the case, and sales specialist Philip Kriener laid out that they essentially had three options:

  1. A cold-start, where most of the infrastructure would be rolled out in case of disaster.
  2. A warm-start, as so-called pilot light installation, where the basic configuration would be replicated and only compute resources needed to be added.
  3. A fail-over solution where all Gold-systems are replicated and continually updated to allow an immediate fail-over.

The Solution

It became apparent that the customer’s representatives where very doubtful how they could convince their board of directors to use a solution that would be quite expensive at face value. “…until we identified together some convincing arguments during the workshop.” explained Kriener, “we compared the current situation with the new concept, where we could show, that the new solution was actually lot more economical than the current one.”  Sandfort added: “the hardware refresh would have been the single most expensive item, even when taking the additional manpower required to realize the failover using VMware Cloud on AWS”. Sandfort explained that the maintenance effort would also be a lot lower using VMware Cloud on AWS than having to maintain the SDDC at the off-site location. 

“The key opener to the concept was the Veeam backup to Amazon S3” explained Sandfort “in case of a disaster at the company headquarters, the backups could quickly restore the SDDC in the VMware Cloud on AWS environment!”

The Results

The customer decided to go with the solution that had the best cost-value ratio: the warm-start solution. The AWS datacenter of choice was Frankfurt and comdivision helped with setting up the SDDC platform, including VPN tunnel, 2-factor authentication and migration of the VMs from the old disaster recovery site. Witness servers where implemented, as well as an RDSH farm with Windows 10 desktops, to be used as emergency jump hosts. Finally, the Veeam backups where established (popuplated with AWS Snowballs) and direct access was prepared to the S3 Buckets. “We are now able to re-establish production within hours by just adding new hosts to the SDDC to cover the compute demand” said the infrastructure admin, and continued “we have prepared the infrastructure, so that after the VMware Cloud on AWS takes over, we can provide 90% of the 640 Desktops to our co-workers within 2-3 days tops!”. “comdivision helped us to understand what we have – sometimes it takes a few pointers and a sounding board to come up with a solution, it was important that Yves and his team did not just created the solution for us, but took us along in the decision-making process!”