Migrating Legacy Databases to the Cloud

SQL-serverLegacy data is often saddled with negative connotations in the enterprise: old, stodgy, kludgy, and obsolete. Valuable nuggets remain in those data repositories, but technology and how these new and old data sources merge continue to transform.

Some organizations with a growing amount of Big Data streaming into it on a second-by-second basis are compelled to upgrade to a new database in the cloud. Older organizations will have a relational database management system, which was once the ideal solution; however, these warehouse-bound repositories cannot keep pace with the volumes of data being produced, which is why IT professionals are opting for open-source, cloud databases.

The goal is having a database that is agile and capable of providing useful information to the business such that they can achieve operational efficiencies or gain a competitive advantage.

“It’s clear that the database is critical to successfully managing the explosion of data,” writes Matt Asay, vice president of corporate strategy at 10gen, developer of MongoDB. “What’s less clear is how to transition from legacy RDBMS to modern NoSQL databases. Successfully migrating from a relational world to a NoSQL world requires careful planning.”

Most often, the migration process will entail working with new applications and equipment, perhaps even new service providers. Establishing good working relationships with these key providers is essential to ensuring a successful migration, whether the process is gradual or accomplished in one big push.

When the scope of the migration project is determined, the database administrator must evaluate the legacy data as it “may not make the cut or be agile enough to survive a migration to an upgraded database or database in the cloud,” notes TrackVia, “but its nature and context remain highly relevant and can influence everyday business activities. In other words, you still gotta use it.”

Data migration is a process that can vary based on the technologies being used because the structure varies significantly. Moving a relational database into an open-source distributed database management system like Apache Cassandra is challenging. Because of the data structure, the migration process entails different procedures than moving to a NoSQL database such as Riak or MongoDB that have key-value stores.

The actual migration process may take several gradual steps. This may include designing new infrastructure and applications. It could be as simple — if upgrading from a Microsoft product such as Access or Excel — as using a tool for quickly synchronizing and/or populating the new database.

The database administrator must also check the database operation in the cloud, examining how and whether it is performing properly. “The process for migrating from an RDBMS to NoSQL is now well-trod, with a great deal of information available online and offline to help with the process,” says Asay.

With the proper strategy in place, you can free your legacy in the cloud and use it to your organization’s advantage.

Image by SQLServer.com.

Contact Us
Vice President and Practice Leader of Oracle Services, Datavail
Patrick’s background includes 15 years of IT experience specializing in database architecture, database administration and performance tuning. He has managed the infrastructure for enterprise database operations of over 300 databases, including several ranging from 10 gigabytes to 80 terabytes. Patrick has designed and developed comprehensive database administration solutions for high performance, reliability and integrity, including backup and recovery, fault-tolerant connectivity, operations and performance monitoring, reporting, automated storage management, BCDR, SOX compliance and Co-Sourcing. A former manager at Level 3 Communications, Patrick has valuable experience in database architecture and corporate data warehousing. Patrick’s hobbies include skiing, Crossfit, hockey and playing with his kids.

Leave a Reply

Your email address will not be published.
Required fields are marked (*).