Search
  • tomcoffing

IBM Takes Over Teradata

Updated: Sep 7, 2021



The Teradata announcement of outsourcing maintenance agreements to IBM is shocking the Teradata community. Is this step a sign of an IBM takeover?

Teradata support employees are holding their breath to the point of turning blue. However, when critical Teradata outages occur, customers have no choice but to hold their breath and turn to Big Blue.

Teradata and IBM have been fierce competitors for decades, so I think this decision Is a Big Mistake (IBM). Do you remember the IBM acquisition of Netezza? How did that work out? IBM drove a great product into the ground.

So why do you outsource maintenance to your number one competitor? Turning over the keys to the castle makes absolutely no sense. Teradata might be saving a few dollars, but customers stand to lose millions. And Teradata's decision to lay off valuable Teradata global support experts is not only risky, and in my opinion, reckless and unethical.

When a large Teradata customer's data warehouse goes down, it costs some customers a million dollars an hour. If the machine is down for a day, it costs 24 million dollars. So time is of the essence, but don't worry, Teradata customer, the company we have been clawing, scratching, and fighting with for four decades, will be there soon to fix it. And we have provided each IBM professional with a few weeks of training classes.

I can only describe the panic in the Teradata community as a migration tsunami.

How do I know about Teradata migrations? I am considered by many to be the top expert on Teradata, so customers trust me to help them migrate.

People in the Teradata community refer to me as Tera-Tom because I have written books on every aspect of Teradata. In addition, I've taught about a thousand classes to large Teradata customers in North America, Africa, India, China, and throughout Europe. I've also trained Teradata Global Support to diagnose and fix Teradata machines. Trust me when I tell you: Teradata machines are the most complex and expensive systems globally, and they are not easy to repair.

Teradata may not want your business, but I sure do. So let me make the migration from Teradata a straightforward and simple process.

I provide a service called Nexus, the only tool that can simultaneously query all systems. But even more impressive is that the Nexus converts and migrates Teradata tables to any other database platform on the market. Even sweeter, customers can use Nexus to join data from Teradata with the systems they are migrating to with the Nexus Federated Query builder. Nexus joins data across platforms similar to Teradata's QueryGrid. The only difference is that Nexus is faster and merges data from any combination of systems.

I have customers right now migrating thousands of tables off of Teradata to a wide variety of cloud platforms like Yellowbrick.

Ten years ago, it became apparent that customers would one day need to access data from dozens of database platforms, maybe even hundreds or thousands of different platforms. In addition, there is too much data coming from other parts of the world that need joining. So, my development team built Nexus to convert, migrate, and join data across all systems, with no technical expertise from the customer required.

Most of my Teradata customers are currently migrating to multiple data warehouse platforms across the AWS, Azure, and Google clouds. For example, I now have Teradata customers using Nexus to migrate to a combination of Snowflake, Amazon Redshift, Aurora, Google BigQuery, Azure Synapse, MySQL, Postgres, and Yellowbrick.

Of all of the systems above, the platform I recommend is Yellowbrick. Why?

The performance of a Yellowbrick system is simply exceptional. The industry-leading price/performance is why I have more migrations to Yellowbrick than any other platform. However, I also look at other critical elements besides price/performance. First and foremost, how fast can a data warehouse load data while also maintaining the concurrency of user queries? Yellowbrick loads data fast while simultaneously delivering subsecond performance to hundreds of users with tables that often have over a billion rows.

While some data warehouses are public cloud-only, Yellowbrick provides the ultimate deployment flexibility, enabling customers to deploy in private, public, hybrid/distributed clouds, and on the edge. Yellowbrick is deployed as an optimized instance in private clouds and/or as SaaS in public clouds with support for AWS, Azure, and Google Cloud Platform.

I was so impressed with Yellowbrick that I directed my Nexus team to automate migrations to Yellowbrick from all data warehouse platforms. And then to ensure that customers can join Yellowbrick data to any systems in which they migrate.

As one major insurance company put it following a Proof of Concept (POC), "Nobody else came within 20% of the speed of Yellowbrick."

Last week, this customer loaded 430 large tables containing 970 GB directly to Yellowbrick with Nexus, and the job took only five minutes to set up.

In a POC against Teradata, a major financial services company and a long-time Teradata customer evaluated its options for data lake augmentation. The company loaded two years of historical data from a MapR-based data lake into each system to compare their performance across a set of 240 queries, with Yellowbrick beating the Teradata system on all but two queries.

I spoke on the condition of anonymity with a long-time Teradata executive, but his first words were, "I left Teradata to work for Oracle." I asked why and he said, "Teradata doesn't know how they perform against other cloud databases."

Maybe the underlying issues of their proprietary hardware are why Teradata is selling off their maintenance to IBM, but ironically, I have many IBM customers migrating to Yellowbrick.

A big reason for the migration of traditional legacy systems is cost savings and performance gains with modern architectures for on-premises solutions. In addition, many companies are migrating to the cloud to reduce the overhead of managing the data on-premises. However, the competition between AWS, Azure, and Google have customers investing in all three clouds, which has given rise to a new model called the distributed cloud model.

The Gartner Top Strategic Technical Trends for 2021 report suggests that the distributed cloud model will emerge to address the explosion of data growth, particularly at the network edge. In addition, companies that adopted a hybrid-cloud strategy see the next evolution in the cloud as a distributed cloud model.

Customers want their data across all public clouds but have one service manage the data across clouds. The potential for saving money is a guarantee because all options favor the customers controlling their destiny, hedging their bets to find the best deals.

Distributed cloud is the first cloud model incorporating the physical location of cloud platforms and the delivery of services in its definition. The customer initially went to the cloud because the physical location was not a consideration. However, a distributed cloud is quite aware of physical location across on-prem, public cloud, hybrid cloud, and edge computing.

The three keys to different physical locations of data are federation, federation, and federation.

A distributed cloud model delivers cloud software and hardware stacks outside of the public cloud provider's data center, providing a mesh of interconnected cloud resources forming a best-of-breed logical cloud.

The two greatest assets a company can provide their prospects are moving data between systems automatically and the capability to federate queries across platforms with no technical knowledge required.

I am thrilled that my instincts were correct and that spending 15 years developing Nexus to query, migrate, and join data across all systems was a visionary idea. As a result, customers can now migrate instantly to and from any database and join data across all systems as fast and efficiently as if the data is one logical distributed cloud system.

I am proud to partner with Yellowbrick to bring customers the next generation data warehouse. At the same time, my Nexus product allows customers to automate the migration and federation between Yellowbrick and all other databases on the market.

Here is a picture of the perfect architecture for a distributed cloud model with crucial elements:

  • Intelligent Nexus Pro Desktops coordinate with Nexus Servers to provide the perfect distributed cloud model.

  • Nexus Pro Desktops allow users to query, move data, or join data across all systems as one logical system.

  • Nexus Servers convert and control the movement between all database platforms over high-speed networks.

  • Users can work from anywhere because they can execute jobs on any Nexus Server.

  • Migrations between any two systems become a simple point-and-click process.

  • The Nexus Super Join Builder Virtualizes the joining of tables and views across all systems.


I am happy to provide a free assessment of your current Teradata environment and prepare a Proof of Concept and a migration plan to help automate your migration process to Yellowbrick.


Thank you,


Tom Coffing

CEO, Coffing Data Warehousing

Direct: 513 300-0341

Website: CoffingDW.com

Email: Tom.Coffing@CoffingDW.com




2,924 views0 comments

Recent Posts

See All