In working with a client, I had the opportunity to evaluate Buldoser from Crown Partners versus OpenMigrate from TSG. Buldozer was used to change OS platforms for a Documentum 5.2 Repository while upgrading it to Documentum 5.3. Buldozer failed to migrate a second Repository due to object level complexities. OpenMigrate was then tested in the migration/upgrade of the complex Repository from Documentum 5.2.5 to new OS hardware and to Documentum 6.5. While both products were designed to assist in the upgrade process, I found they work in substantially different ways.
Crown Partner’s Buldoser – A folder Approach
- Relies on building a bulk export cache of information on the file system and in memory before importing it into the target system. For my client, this caused issues when moving a large batch of documents. Clients should verify performance on a sample set to determine file system and memory requirements.
- Performance – While quick at first, the complex folder structure in the client’s Repository caused Buldozer’s performance to degrade considerably. After a period of time it was taking several minutes to migrate a single document.
- Service/Support – For my client, as a purchased product, Buldozer could not migrate the more complex of the Repositories. Even with support and service updates it was inadequate.
The Buldozer product suite is available from EMC as a standard item on the EMC Documentum price list.
First a disclaimer, as a TSG employee I have had some input into the design of OpenMigrate but the core approach was completed before I joined. That being said, the use of a node versus a folder approach was the substantial difference I found. Specifically,
- OpenMigrate relies on an alternative approach by migrating documents individually. The product uses the concept of a migration node. Each node represents an object such as a document, folder, or user.
- The set of nodes in a migration queue is built by one or more queries into a document management system. The queue can also be built from a list of documents in an Excel file. Multiple batches of OpenMigrate can be run at the same time either working from different queues in memory or a shared queue stored in a database table.
- By operating on each node individually, OpenMigrate can scale up or down to make the most efficient use of available resources. For example, OpenMigrate is multi-thread configurable in three areas: connecting to the data source, connecting to the data target, and transforming, or mapping, data values.
As a free software product, you can try out OpenMigrate’s sample configurations in about 30 minutes. The Open Migrate Binary is downloadable at http://www.tsgrp.com/Open_Source/OpenMigrate/open-migrate.jsp.
Please comment your thoughts.