1. Locate the relevant row(s) based on the search conditions specified in the update statement.
2. Lock the row(s) to prevent concurrent modifications from interfering with the update.
3. Update the data in the row(s) as specified in the update statement.
4. Update the non-clustered index to reflect the changes made to the data.

Updating the non-clustered index involves a few steps:

1. Identify the affected index entries: The non-clustered index contains pointers to the actual data rows. The update operation needs to identify which index entries are affected by the update.
2. Modify the affected index entries: Once the relevant index entries are identified, they need to be modified to reflect the changes made to the data. This may involve updating the key values, adding or removing entries, or adjusting the position of entries within the index structure.
3. Maintain index integrity: The non-clustered index needs to maintain its integrity after the update. This may involve reorganizing the index structure or rebuilding parts of it to ensure efficient data access.
4. Release the locks: Once the update and index modifications are complete, the locks held on the affected rows can be released, allowing other transactions to access and modify the updated data.

It’s worth noting that updating a table with a non-clustered index can have an impact on performance, as it involves additional steps to update both the data and the index. The extent of this impact depends on the size of the table, the complexity of the non-clustered index, and the number of rows being updated.