Exploding demand for increasingly rich data types is putting significant pressure on storage infrastructure at all levels of the organization. Staying ahead of the curve, however, involves more than simply throwing in a few new drives and switching them on. Upgrade projects should be carefully planned and executed to avoid data availability disruptions.
Although storage upgrades often involve new hardware, moving your existing data onto it may not always be the optimal choice. Minimizing the impact on what’s already stored is a crucial component of any upgrade project, says Satish Joshi, executive vice president of Patni Computer Systems .
“Data migration can be a challenge and a nightmare,” Joshi says. “Ideally, one should aim for in-place upgrades where they can upgrade the storage systems for higher capacity and performance without moving the data.”
Talk It Out
Although keeping data where it is may be preferable, it may not always be feasible or even possible. Migration best practices can help reduce the resulting migration-related risk. To avoid migration disasters, Joe Hayes, director of storage for QTS (Quality Technology Services; , says IT should communicate early, often, and thoroughly.
An ideal first step is identifying everyone who will be involved in the process.
These could include staff in charge of such resources as OS management, to ensure that updates and changes are supported in the operating environment; storage, to cover all application- and site-specific needs; backup, to ensure that backups are taken before migration and to ensure that they work afterward; applications, to test post-migration application functionality; disaster recovery, to update failover processes; and network, to ensure that the underlying infrastructure remains capable of supporting the new storage environment.
Once the groups have been identified and contacted, Hayes recommends assigning a point person from each one and scheduling a meeting with these key resources.
“This verifies that everyone involved knows what is going on and should hopefully prevent the ‘Hey, nobody told me about this!’ scenario from happening,” he says.
The meeting should focus on reviewing the migration plan, soliciting everyone’s input on where possible changes can be made, and confirming testing criteria for determining whether the migration is successful. The meeting should also confirm who will be part of the migration groups for each affected area and when the migrations themselves will take place.
Have A Storage Technology Roadmap
Joshi says a range of new technology in the storage space is making upgrade projects easier than ever to plan and execute. He says upgrade projects provide IT an ideal opportunity to review and optimize the technology mix. A range of emerging architectural choices can minimize risk during and following implementation:
Tiered storage. Tiering your storage lets you use different technologies for different needs. For example, SATA drives can provide a low-cost, high-capacity layer, while more expensive Fibre Channel-based storage can address higher performance needs. Flash memory can be used tactically for apps that can differentially benefit from it.
Automated tiering. Joshi says new technologies support automatic storage tiering. This allows preset service levels to be assigned to specific applications, which in turn lets IT dynamically move data across tiers to optimize both performance and cost.
Storage virtualization. Long used for servers, virtualization is increasingly being leveraged for storage, as well. At this level, it supports data consolidation and provisioning as well as dynamic workload movement. Similarly, thin provisioning at the storage layer allows efficient, application-aware storage allocation as data needs grow.
Additionally, methodical planning is critical to a successful storage upgrade, Hayes says. A big bang transition can open the door to spectacular failure and should be avoided at all costs.
“Moving in groups helps to limit the exposure if something goes wrong and will help to keep the process moving more smoothly,” Hayes says. “The worst thing that can happen is trying to do too much at one time and an unexpected delay hits and everyone becomes unhappy.”
SANs Demand Special Attention
For SAN-based solutions, IT must also ensure that components such as operating systems, applications, and drivers are supported in the target environment. When migrating SAN data, Hayes says, choosing the right replication tool is key.
There are two main categories. Host-based tools specifically support a given OS and storage array. As a result, they must be installed and maintained, and IT might require numerous tools to support each OS.
SAN-based tools, on the other hand, are less intrusive than host-based alternatives, as they can be set up to run in the background. They can also support no-downtime migrations, but they can be more expensive.
In all cases when buying a new array, IT should ask the vendor to provide migration tools at no charge. Where applicable, the migration plan should also incorporate external integration components such as replication and SAN optimizers.
Backups Matter, Too
Storage upgrade best practices extend to the backup environment, as well. Patni’s Joshi says continued growth in data often outpaces that of traditional backup methods.
The net result: Backup windows eventually become too narrow to accommodate the larger pools of data. Newer backup solutions have the speed and capacity to keep pace, but only if they are implemented in lockstep with the storage environment upgrade.
“The new storage systems that you upgrade should provide disk-to-disk backup solutions with deduplication capabilities at the source so data can be easily moved to DR locations,” Joshi says.
“This will not only bring down tremendous overheads of tape management and DR site shipping, but significantly enhance data restores if and when required.”
Source:http://www.processor.com/editorial/article.asp?article=articles%2Fp3218%2F22p18%2F22p18.asp&guid=&searchtype=&WordList=&bJumpTo=True
