Live Data Center Migration

Summary

Initiated by consolidation measures and the need for centralization, this client’s data center migration consisted of moving a suite of applications spanning ten virtual machines across the country from a private data center to a federal facility. Minimizing end-user disruption was of the highest importance—as an agency committed to advancing the safe transportation of energy and other hazardous materials, our client’s workforce is more impacted than most by obstacles to necessary information. In full support of their mission and the high ROI a successful relocation would bring them, CSNW leveraged its extensive VMware and Windows Server operations experience to perform this migration seamlessly.

Goal

The potentially hazardous nature of our client’s daily operations affords them an incisive perspective on data: there are few assets more valuable, but when mishandled, data becomes an obstacle. In consolidating resources, our client aimed to simplify disaster recovery, centralize management, and improve business continuity. At CSNW, our goals for the data center migration also included minimal end-user disruption and secure, efficient service delivery.

Work Performed

In consultation with IT and infrastructure support personnel at both data centers, we began by creating a generalized O&M document for server relocation. As this was the first time our client had attempted such a move, we opted to test the O&M document on a non-production exemplar server as a means to refining both the process and the documentation. This early testing phase validated our initial plan to deploy a parallel architecture while maintaining the original in production rather than shut down the existing servers and send everything over to the new data center. After identifying owners and stakeholders for each application and establishing a communication plan for each, we used a tiered syncing approach to transfer the data continually while the applications were running.

After 90 days of successful running and validation of O&M procedures at the target location, we scheduled and performed final cut-over, decommissioned the virtual servers at the original data center and deleted them from the storage array per federal requirements. To make sure all client data left at the original location was destroyed, we retired and wiped the storage array on which the old hosted servers were running, as well as on the backup storage.

With our client’s safety mission top of mind throughout, we were able to anticipate many potential modes of failure and design the necessary safeguards to mitigate any resulting issues. We have a library of scripts built on a common framework that enables us to quickly solve a variety of replication and migration problems that would traditionally be tough to troubleshoot.

Results

In addition to improved business continuity and consolidation, other deliverables include:

1. Due to the small amount of information transferred in the final sync and the minimal delta for data sync during off-peak hours, downtime was measured in seconds. No users were impacted.
2. In accordance with SP 800-115 Technical Guide to Information Security Testing and Assessment and SP 800-34 Rev. 1 Contingency Planning Guide for Federal Information Systems, DR and contingency plans are updated to reflect the system architecture changes.
3. Security assessment documentation on SP 800-53 Rev. 4 Security Controls (specifically Physical Security and Maintenance) are updated to reflect new application environments.
4. This data center migration also presented the opportunity to improve server efficiency via an OS upgrade.

While a cross-country data center move is not a routine operation, this transition was made in such a way that the routines of the people that relied on it were not disrupted. The client’s fiduciary and regulatory constraints were met while at the same time, the important safety inspection work that this system supports was not disrupted.