With 1000+ servers spread throughout France, the main goal of Transdev’s Disaster Recovery project was to reliably secure the servers running the most critical workloads in corporate data centers, at the lowest possible cost. After years of using a traditional on-premises disaster recovery strategy, Transdev decided it was time to look into leveraging the public cloud for disaster recovery. Several factors went into this decision.
First, they were attracted by the lowered disaster recovery infrastructure costs afforded by the pay-as-you-go cloud model. Second, since they had an AWS migration project underway, they wanted a disaster recovery solution using Amazon Web Services (AWS) as their disaster recovery target site. Finally, Transdev had to resolve the compatibility issues they were experiencing between their primary on-premises data center and their target on-premises disaster recovery site, since the primary site had been refreshed at the infrastructure level while the target had not.
As Matthieu Traverse, Lead IT Architect at Transdev, describes, “We identified that the cloud was the right strategy for the evolution of our internal IT infrastructure, and since we needed to refresh the disaster recovery infrastructure, we took the opportunity to build it directly in the cloud… We wanted to create something that would not lock us into the old world.” As a result, Transdev turned to Gekko for disaster recovery expertise, and to help define, orchestrate, and build a new cloud-based disaster recovery strategy. Traverse worked with Julian Favre, co-founder and managing director of Gekko, to find a disaster recovery solution that would protect Transdev’s IT environment.
Transdev and Gekko reached various dead ends when seeking solutions that could effectively shift their disaster recovery into AWS. Although they performed proof of concepts and demos with several different companies, none were able to provide an AWS-integrated solution for their complex Microsoft workloads, or could only do so with unpredictable performance. Transdev came close to giving up on cloud-based disaster recovery entirely, until AWS recommended CloudEndure Disaster Recovery to Gekko. As Traverse explains, “We tried several solutions but did not find anything that we were comfortable with, so we were about to rebuild on-premises. Then Gekko came to us with CloudEndure and the choice was really easy to make.”
Transdev chose the toughest and most complicated use case for proof of concept (POC) testing: spinning up their SAP BusinessObjects Financial Consolidation (SAP BFC) application running on Microsoft Cluster Server (MSCS) infrastructure, to see if it could securely failover to AWS. “We thought that if we managed to validate this most complicated application and most complicated infrastructure, we would be able to do so with every application,” explained Traverse. AWS is not typically considered a disaster recovery target candidate for MS clusters such as SQL Server clusters, but CloudEndure technology is able to reliably perform this replication and recovery using proprietary Machine Conversion technology, which is effective for any application. In addition, part of the POC included Gekko using CloudEndure’s open APIs to customize the disaster recovery solution for Transdev. As Favre explained, “We developed a new tool based on the CloudEndure API in order to orchestrate, manage, and operate all of the disaster recovery services within the defined recovery objectives.” Anything that can be manually configured in the CloudEndure UI can be automated using open CloudEndure APIs, allowing Gekko to customize disaster recovery solutions for Trandev and other customers.
The POC with CloudEndure was successfully concluded within one month, after the technology proved its effectiveness in implementing Transdev’s most complicated use case, supporting Gekko’s additional tooling, and deploying the first 20 machines for disaster recovery.
From deployment of the first twenty servers during the POC, through the completion of the disaster recovery project, there were no performance disruptions at any stage of deployment, including during agent installation and initial block-level replication of file server workloads. Traverse added that “Deployment of the CloudEndure solution was really quick,” as full deployment was completed within four months. Transdev’s IT environment is entirely built on Microsoft Windows infrastructure. The disaster recovery project included Microsoft products such as Active Directory, Cluster Servers, and several SQL Server databases. In addition to the SAP BFC application mentioned above as part of the POC, several lesser-known niche applications were also included in the project.
After completing the project, Transdev was pleased with the cost reduction achieved by building their disaster recovery target in AWS with CloudEndure and Gekko, versus meeting the same requirements through refreshing and building an on-premises solution. Regarding the planned costs of each option, Traverse stated that “We estimate that building our disaster recovery in the cloud was 73% cheaper than if we built the same solution on-premises.”
Now that Transdev’s critical workloads are protected, they plan to perform two related projects this year. The first is to execute an additional annual set of disaster recovery tests, which CloudEndure easily enables at the click of a button, with no performance impact. The second is to deploy new versions of some of their protected applications, and ensure that seamless machine replication and disaster recovery protection continues after deployment.
Thanks to the success of this project, Transdev is currently implementing an Gekko designed migration of their entire production environment to AWS, leveraging CloudEndure technology for aspects of this process as well.