Financial Case Study
|Customer Type||Major Global Accounting Firm|
|Source Storage||EMC VNX|
|Target Storage||HPE 3PAR|
|Number of Hosts||1488 Hosts, Many Clustered|
|Host Types||Windows, Linux|
|Number of DMS and model||2x DMS2000P for each VNX|
|Insertion Method||Logical Storage-Side|
No downtime available until months after migration deadline
Limited information on existing environment
No changes to the source VNX storage were allowed due to known risks
Short implementation timeframe
Using TDI patented technology, DMS was inserted into path totally transparent to the hosts, FC switches, and the VNX storage
DMS auto-discovered all hosts, source LUNS, and paths
Maintained compatibility with PowerPath even after pMotion and I/O were going solely to the 3PAR array
Maintained compatibility for boot from SAN
Eliminated risk of changes and application outage that are usually associated with using standard migration tools (host or appliance)
Delivered on-time migration with predictability with full control over infrastructure and migration process
pMotion allows old storage to be removed without affecting the host by emulating the old storage using the new storage
No dependencies on other software, OS, storage array, applications versioning, or SAN array changes—allowed the migration to move forward without requiring remediation until final cutover
A major global accounting firm recently faced a critical storage refresh challenge. It needed to immediately upgrade an aging storage infrastructure to avoid the risk of outage and to correct performance issues. Unfortunately, the downtime that is typically required for a storage migration was not available until many months later. The company needed a high performance, real-time, and non-disruptive method to migrate from their legacy EMC VNX to their new storage platform choice, HPE 3PAR FlexCapacity, but also needed to defer the cutover downtime for several months.
The most difficult part of both planning and executing any data migration is the final server cutover to the new storage. The downtime typically required for this migration must generally be scheduled far ahead of time. In this case, the old storage was putting the application service levels required at risk, which required immediate action. The planning and execution efforts using previous methods would not only add to the time, resources, and risks required to move the data and complete the data migration but would also require immediate downtime.
What if there was a solution that allowed the business to retire its old storage systems and defer the final cutover downtime for several months? What if the solution could be deployed transparently, at high speed, and eliminate the need for any host, FC/SAN, or source storage configuration changes—processes which could result in months of risk analysis and change approval?
Holding on to old production storage can incur hefty leasing extension fees, can pose unacceptable risk to data availability, and can even result in data loss. CDS premier partner realized the client needed a solution that would allow them to quickly retire their old VNX storage assets and migrate to 3PAR while deferring cutover downtime. HPE experts were quick to recommend Cirrus Data’s Data Migration Server (DMS) with its unique pMotion capabilities which allows using the new storage to stand-in for the old storage indefinitely—before final cutover. pMotion is particularly useful in scenarios when source storage is failing, lease expiration is imminent, or the maintenance contract is expiring and downtime for cutover is not available in required service windows.
This is how pMotion works: The DMS appliance pair is placed into the FC fabric using Cirrus Data’s patented TDI technology, which allows SAN insertion without downtime and without making any changes to any host, the FC switch zones, or the production storage. Upon insertion, the source LUNs are automatically discovered and new LUNs are created automatically using the DMS storage plugin for the destination storage and paired with each source LUN. Migration sessions are then created and enabled based on the priority established by the customer. In this instance, migration performance was over 1GB/sec. Once the initial copy is completed, the pMotion feature is triggered, effectively retiring the old source LUNs and redirecting all the read/writes to the new storage.
DMS emulates the decommissioned storage; the application servers and FC switches still “see” the old storage and datapaths, avoiding the need for downtime to reconfigure each host to see the new storage. The pMotion migration process is a totally transparent action and does not impact the production environment. This solution makes it possible to defer the final server cutover for weeks or even months, as long as the DMS appliances continue to be deployed. For this project, the pMotion functionality was imperative and allowed the client to adhere to its own timeline for scheduled application downtime that was available months later.
Two other product features supplied by DMS were key to the successful completion of the project:
- DMS Auto-discovery identified storage LUN-to-host mapping, which was vital due to limited documentation on source storage LUN configuration
- DMS Intelligent QoS (iQoS) monitored actual load on each LUN, allowing maximum use of all available bandwidth for copying data without impacting production I/O quality
With DMS’ simple and intuitive operation, HPE and its client were ready to migrate DMS in just hours (not days or weeks). After appliance insertion, the actual migration and cutover were performed remotedly. CDS technicians worked with the client’s IT staff and HPE migration experts to provide a seamless migration from start to finish. This deployment proved that DMS provides the flexibility of a single solution that can meet any client’s migration requirements with or without the use of pMotion, while reducing or eliminating the complexity and risks of moving from any legacy storage to any next-generation storage, regardless of storage vendor.