Different storage repositories will see different migration performance based on the use case. This page provides the average migration rate expected by repository.  The rates vary depending on whether the repository is the source or the destination.

** The NFS averages are completely dependent on the bandwidth from or to the NFS volume. This can vary greatly depending on the bandwidth.

Factors That Determine Performance

• The numbers above are an average over the full life cycle of a migration. Performance will have numerous peaks and valleys over time.

• The first run of migration jobs will see the best performance. DryvIQ will be running at capacity to migrate as much content as possible. On delta runs, there is less content to migrate, so performance appears slower. This is actually not accurate; it just has less work to do.

• The data makeup is critical to performance. A large number of files will take longer to migrate than a smaller number of files. This is true even if the small number of files are the same size or larger in total volume. Also, a concentration of data under a single user or location can greatly degrade performance.

• All cloud platforms have rate limiting to preserve tenant health. End users are prioritized over backend processing. Tenants are designed to slow down processing like migrations to maximize the end users' experience.

• Repositories like Box provide a fixed number of API calls over a designated period of time. This can vary by customer as these are part of the contact the individual customer has with the repository vendor.  Customers can pay for more API calls, but it can be expensive.

• The amount of hardware used for the DryvIQ environment can affect performance. When maximizing performance, the environment should be scaled up to run as fast as possible while staying under the rate limiting settings for the tenant.