Migration Challenges: Cloud Vs On-premises Migration

The traditional on-premises migration involves transferring large datasets or files from one server to another, typically before or during the live migration window. Data from various legacy systems is extracted, transformed, cleansed, and loaded into a staging table. The data is then loaded directly onto the underlying base tables of the target system through intermediary staging databases. For high throughput, the target system infrastructure is scaled horizontally and vertically. In this scenario, LAN connectivity plays a significant role in ensuring smooth data transfer between systems.

When migrating to a cloud-based SaaS/PaaS target system such as Google Cloud Platform (GCP), SugarCRM or Salesforce, data must be transferred through the API layer adhering to the multitenant environment governor limits. Even the Bulk API approach has governor limits that allow only 5,000 batches in a 24-hour rolling window, each having a maximum of 10,000 records . However, the actual throughput will be much lower due to custom code and various other governor limits. It is essential to evaluate each cloud application to assess the available transaction and governor limits and understand how they are consumed during migration.

Further, several out-of-the-box (OOTB) industry cloud and managed package triggers and automations are activated in the data layer. These must be considered when calculating the overall processing time. Network latency can also impact cloud migrations, as all data transfer occurs over the internet.