Data migration used to be a relatively simple task: an IT team would identify a new location, make a copy of the data in question, store it on a larger drive, and then change the path to access it. However, in recent years cloud servers have become increasingly viable options for data storage, but, because of the need to avoid downtime at all costs, data migrations on the cloud can be complex.
Complex, but not impossible: there are several different ways to migrate data from onsite servers to a cloud environment (or even from one cloud server to another). The degree of security and efficiency varies based on the migration method used.
Selecting Your New Server
After you’ve made the decision to migrate data, you’ll need to decide where the data will be stored. This may sound simple enough, but many companies who initially moved data from an onsite server to a cloud environment are now considering migrating data from one cloud to another. The reason? Not all clouds guarantee the same speed, uptime, and data recovery features. Check the Service Level Agreement (SLA) to ensure your data will be both secure and accessible for years to come — otherwise, you may find yourself migrating again in the near future.
Determine Your Migration Method
Depending on the size of your data migration, you’ll need to decide whether to transfer your data over the internet or ship it to an off-site facility. Internet transfer over a Wide Area Network (WAN) may seem like the obvious choice: under optimal conditions with an extremely fast connection, it takes about 2 hours to move 1TB of data. That said, you may not always have optimal conditions, and most 1TB data transfers take closer to 5 hours. If you’re a small organization with just a few terabytes of data, this may not be a problem; however, for larger organizations with hundreds of terabytes (or even petabytes) of data, internet transfers may be neither efficient nor feasible. Issues can also arise if connection to the WAN is lost, possibly compromising the integrity of your data.
When a large amount of data is being migrated, physical moves are often more effective. In scenarios like these, data is usually moved to a backup location on site, then shipped to a remote data center. While a full migration of petabytes worth of data can take months, with physical moves, this can be done in waves. Migrating data over time can ease the transition for employees and limit downtime.
Primary or Backup Storage?
You’ll also need to determine the primary purpose of your new storage location: will it be the main source of data storage, or will it be used as a backup location? If the new server will be used as a primary location, it can be written to simultaneously with the old server until a full migration has been completed. This ensures that data syncs between new and old servers, preventing unnecessary loss. Once the two locations are identical, the old server location can be used as a backup or deactivated completely.
The simultaneous writing method is also used with databases — but this requires a few additional steps after the initial migration has been completed. When migrating data from a public-facing application from a database, you’ll need to ensure your DNS move is transparent, limiting latency and keeping traffic moving. It’s important to check that custom code or scripts have also migrated properly, as well as ensuring all methods of calling data from and writing data to the database are correct. Testing applications and stress testing the database in a sandbox environment is also critical before everything goes live.
Although migrating to the cloud (or between cloud servers) can seem daunting at first, more and more companies take this important step every day. By the end of 2017, public cloud services will be a $246.8B industry. Whether your company plans to move to a full cloud architecture or a hybrid cloud setup, the time to transition is now.