By Brian Peterson, Contributor
It seems like everyone is buying storage. The combination of huge data growth rates (more than 50% a year at most IT shops), dropping hardware costs (by at least 20% annually) and steady maintenance costs is making it more cost-effective to replace storage arrays with higher-capacity ones than add to an existing array.
A natural play for integrators is to add new storage frames to support data growth while also consolidating and eliminating older, more expensive storage subsystems. It lowers customer operating costs and at the same time increases reseller revenue. However, customers often shy away from consolidation because they fear large, complex and disruptive data migrations. Successful storage integrators can address this concern by offering data migration services based on solid methodology and great tools.
Here's my unscientific take on the best data migration tool in each of five categories. To be considered, the tool must be optimized for one-time data relocation from one storage device to another, with an emphasis on heterogeneous replication. Special emphasis is given to data migration tools that don't have to become a permanent part of the infrastructure.
Category 1: Host-based file-level migration
And the winner is: rsync. This open source tool has been around for a long time and distinguishes itself by being very simple, yet powerful, and totally host- and storage-agnostic. rsync is very flexible and can be adapted to almost every data migration need, but it shines especially brightly with largely static unstructured content.
Category 2: Host-based block-level migration
With large structured files like databases, block-level migration tools make the most sense. I'm going to cop out here and not name a specific tool but instead a spectrum of tools: Host-based volume managers are often overlooked as a data migration tool, yet they provide a powerful way to nondisruptively migrate data from one storage array to another. Most operating systems already have a capable volume manager that is heterogeneous and already installed.
Category 3: Network-based file-level migration
Sometimes the data migration simply can't be done on the host. This is especially true when a lot of hosts access the same data, as happens with NAS arrays. The winner in this category is EMC's Rainfinity. This NAS virtualization appliance can be inserted into the data path between the servers and the storage array, orchestrate nondisruptive migrations and then slip quietly back out of the data path.
Category 4: Network-based block-level migration
Storage area networks (SANs) were once just a place to route servers to disk. These days they have become much more sophisticated, and intelligent fabric services are not only possible, they are commonplace. Brocade's Data Migration Manager (DMM) is a SAN-based heterogeneous data migration tool that leverages the company's AP7600 intelligent SAN device. Migrating logical unit numbers (LUNs) from one storage array to the next is possible with several SAN-based tools, but this one is different because it moves the data online and doesn't have to take control of the LUNs.
Category 5: Array-based block-level migration
It is nearly impossible for an array-based data migration tool to be heterogeneous -- unless the array itself is heterogeneous. Hitachi Data Systems' (HDS) Universal Replicator can migrate data that is both internal and external to HDS Universal Storage Platform (USP) arrays. This type of replication works great if the customer already has or is moving toward a HDS USP array and the hosts cannot support the workload required to move the data.
There you have it, five storage replication tools in five separate categories. Solution providers who can wrap data migration services around a few of these tools will bring more value to their customers and more revenue to themselves.
About the author
Brian Peterson is an independent IT infrastructure analyst, with a background in enterprise storage and open systems computing platforms. A recognized expert in his field, Brian has held positions of responsibility on both the supplier and customer sides of IT.
This was first published in August 2008