Help Center/ Server Migration Service/ FAQs/ Migration Duration/ What Factors Affect the Migration Speed?
Updated on 2024-08-15 GMT+08:00

What Factors Affect the Migration Speed?

The following table lists the factors that may affect the migration speed and duration.

OS

Factor

Description

None

CPU or memory usage

The migration occupies some memory and CPU resources of the source server. The usage varies depending on the source conditions. Before the migration, ensure that both the CPU usage and memory usage of the source server are not higher than 75%, and the available memory is at least 520 MB.

Network bandwidth

The network latency between the source and target servers must not be too high. The migration speed depends on the source bandwidth and the target bandwidth, whichever is smaller. For details about network requirements, see How Do I Set Up a Secure Migration Network for Using SMS?

Windows

Disk fragments

SMS supports only block-level migration for Windows servers. Only valid blocks are migrated. A large number of disk fragments will be generated on disks in daily use. It takes time to identify valid blocks among these fragments.

Linux

  • Large files
  • Too many small files

For a file-level Linux migration, the migration speed will be affected if

  • There are files larger than 2 GB.
  • There are a large number of small files, such as, those smaller than 20 KB.

The following table lists the factors that may affect the synchronization duration.

OS

Factor

Description

None

  • Too much new data on the source server
  • Too much changed data on the source server

During an incremental synchronization, if a large amount of data is newly generated or changed on the source server, the synchronization takes a long time.

Windows

Too many file fragments

During a synchronization, if a large number of file fragments are generated on the source server, the synchronization takes a long time.

Linux

Large sparse files

During a synchronization, the system scans but does not migrate sparse files. If there are large sparse files on the source server, the synchronization takes a long time.