In part 1 of this series I outlined the considerations and process required to perform nondisruptive upgrades and migration. In part 2 I describe our most challenging customer case study: direct cloud-to-cloud bulk migration of data between two providers and changes to the protection level of data within a provider. Both cases involved moving large volumes of data outside the customer’s infrastructure and without any disruption to the end-users.
Process in action
Let’s take a look at the entire process for one of our customers that upgraded. This customer had three filers on three different accounts. They had a number of volumes on Amazon S3 and Rackspace Cloud Files and some of their Amazon S3 volumes were set for reduced redundancy (RRS). For us to offer the new service to them we needed to get them upgraded from RRS and all their data resident in Amazon S3 - and we needed to do it without interruption to their business. The customer had twelve terabytes of data with Nasuni at the time of the upgrade. This customer also purchased some hardware from Nasuni to improve performance as part of the upgrade. In order to bring all of this customer’s data under the new Nasuni SLA-backed storage service we had to:
- Copy all of the data from Rackspace to Amazon using cloud compute so as to avoid overloading the customer’s own network. According to our security model, the copy operation involved moving encrypted objects from one CSP to another without ever having visibility to the data.
- Ask the customer to DR his three existing virtual filers to three new filers two virtual filers and one hardware filer. The customer experienced a few minutes of outage for each DR as they moved from one product to another.
- Nasuni then merged the three accounts for their three filers into a single account for the new service
- At this point, the customer suspended snapshots. Nasuni did one last copy of the encrypted data (again using cloud compute resources), and then switched the providers for their volumes, and the customer resumed snapshots.
The only downtime to move from one product to a completely different product and shift terabytes of data around in the service was a few minutes for each filer for the DR. The customer is now at fifteen terabytes and has better performance, better protection, guarantees around the service and they can access their data globally.
In our customer’s own words:
“Now it is truly a "set it and forget it" type of service. In addition, the multi-site capability is really great as it no longer requires our users and sister companies to have to VPN to get to our filer's data. Not only this makes data access more convenient for our remote users but it also greatly improved data retrieval when compared to going through a VPN tunnel as was the case before.”
Storage services are no different from traditional storage in that what customers are ultimately purchasing is a commitment from the vendor to safeguard their data. However, enterprise storage services must take that commitment well beyond the physical storage controller. Our job at Nasuni is to ensure that customers’ data is safe and always available regardless of what happens.