Supernova Award Category
The Problem
Core Digital needed to set up a cost-effective disaster recovery system that minimized down time.
Core Digital wanted to avoid deploying and maintaining a separate physical disaster recovery (DR) system for its on-premises Teradata data warehouse.
In 2014 the company embraced a private-cloud strategy, moving to increasingly virtualized (VMware) application workloads running on industry standard (Cisco UCS) hardware running in a hosted data center in California. The company has a backup data center in Colorado that provides virtualized infrastructure as a service as well as hosting services. Until 2015, the company housed Teradata capacity exclusively inside the California data center. The Teradata footprint included a data warehouse appliance, used for production, and older-generation racks used for disaster recovery. Lacking a geographically isolated Teradata DR instance, the company had no fail-over options in the event that connectivity was lost or a major disaster struck the California data center. Tape backups were maintained, but business continuity was at risk.
A disaster recovery system project posed the following challenges:
- Cost: The company wanted to minimize capital expenditures, avoiding buying, deploying and maintaining a second system.
- Down time: Faced power, space, deployment and ongoing administrative and maintenance burdens tied to running a second system.
The Solution
Core Digital decided against the capital expenditure and delivery, deployment and testing delays required by a physical deployment and deployed Teradata Cloud Disaster Recovery System in April 2015.
The results
Core Digital did a proof-of-concept project of the Teradata Cloud Disaster Recovery System in April 2015. Having satisfied all performance and data-replication tests, the firm signed a contract for Teradata Cloud in June and the DR instance was operational by July.
Hybrid deployment combining on-premises production systems with Teradata Cloud had a positive impact at Core Digital.
Avoided capital investment and reduced costs. Core Digital not only avoided a large, up-front investment in a second physical appliance, it also took advantage of cloud elasticity, opting to use two cloud computing units to back up a four-node appliance. Cloud capacity can be added as needed. Finally, legacy on-premises systems used for DR were retired, eliminating six-figure maintenance costs.
Reduced administrative burdens. Teradata Cloud takes care of backup and routine software maintenance, such as patches and updates to the software behind cloud computing units. Core Digital avoided additional system administration burdens.
Mitigated risk and improved flexibility. Core Digital previously lacked a geo-isolated DR center, so it risked outages in the event of lost connectivity or a major disaster at its production data center. That risk is gone now that it can switch to a different site and it can also rely on the DR instance for uninterrupted operation in the event of planned outages at the production data center.
Metrics
- Cost effective disaster recovery in less than two months. The project ensures fail-over even in the event of major disaster. Proof of concept commissioned in April 2015, and instance operational by June 2015.
- Avoided large capital investment and ongoing power, space, administrative and maintenance costs of on-premises system.
- Passive use of Teradata Cloud capacity minimizes costs. Cloud is less expensive than the full Teradata Database as a Service (TDaaS) geared to active use. A single cloud computing unit (CCU) of DR, is $2,850/month vs. $3,700/month for a single CCU of TDaaS (with both figures based on 3-year commitments).
- Cloud scaling minimizes costs. With the cloud deployment Core Digital only pays for the capacity needed rather than matching the capacity of the on-premises system. “We have a four-node appliance, but we’re using less than half of its capacity,” Jean explained. “In the cloud, we’re able to replicate and get acceptable performance with just two CCUs... When it’s in the cloud, you can replicate to a scaled-down version of what you have and get more resources when you need them, if you need them.”
- Cost savings with Teradata Cloud-provided VPN connectivity. Core Digital stuck with standard, Teradata Cloud-provided VPN connectivity (rather than dedicated lines) to its production data center in California and an additional, site-to-site VPN connection to its DR center in Colorado.
- Faster time to recovery in the event of a disaster.
The Technology
- Teradata Cloud Disaster Recovery System (passive)
- Teradata Cloud Teradata Database as a Service
Disruptive Factor
Using disaster recovery as a starting point proved the value of hybrid cloud to the rest of the organization. As explained above, Core Digital left their production Teradata systems on premises while moving DR to Teradata Cloud.
This deployment solved problems and yielded expected and unexpected benefits. As a result of the successful deployment of the Disaster Recovery system, Core Digital now perceives hybrid deployment as an invaluable aspect of its data warehousing strategy. What’s more, they’re investigating additional hybrid use cases, starting with data exploration and predictive analysis in the cloud.
Shining Moment
Thanks to the DR project, the flexibility of cloud capacity is now top-of-mind at Core Digital. Core Digital periodically does ad hoc predictions that require deep historical analysis, an activity that affects production performance. Aware that the organization could move these performance-draining workloads off of on-premises,
Core Digital is considering moving that work to the cloud.
