Teradata invented parallel processing and sold its first system to Wells Fargo back in 1988. I was one of the first Teradata teachers and taught at places like Walmart, AT&T, Bank of America, Southwest Airlines, and Blue Cross Blue Shield. These companies, guided by Teradata technology, grew to become icons in their respective industries.
Although Teradata has been around for 30 years, if you are a company that needs an on-premises enterprise data warehouse, then Teradata is a great choice. Does Teradata still have the best data warehouse platform in the industry? Absolutely yes! Without a doubt.
How would I know? I have taught 1,000 classes and written over 85 books on all aspects of Teradata, as well as all major systems. I have written 96% of the books under the category of computer engineering, so I am the top authority in the world on database platforms.
What is so unique about Teradata is its ability for active data warehousing. Because Teradata scales linearly to a massive size, a company might have a table with a trillion rows, but if you use a WHERE clause in your SQL on the table’s Primary Index, the query will return in one second. So, you get the best of both worlds with parallel processing of large tables for data discovery and analytics, as well as sub-second tactical queries, and everything in between.
You can even load data in real-time as users continue to query. They have the largest and most sophisticated data warehouse customers in the world, and their customers love the technology.
Teradata’s incredible speed comes from a combination of their database design, hardware configuration, and especially from their BYNET technology. It is the BYNET that connects the nodes and provides intelligent speed beyond all competitors.
And Teradata has learned from competitors such as Netezza, Oracle, Kognitio, and Redshift.
Teradata combines the techniques from all vendors. All vendors use parallel processing, but Kognitio and SAP HANA are in-memory solutions, which makes for sub-second querying. Some vendors store data in a columnar format, such as Amazon Redshift, which is perfect for analytics and ad hoc queries. But nobody does it all, as well, as Teradata.
Teradata tables each have a primary index(s), which is similar, but faster than a distribution key on other systems. They have up to 32-secondary indexes per table, and you can partition a large table for speed on range queries. Teradata can store data in row blocks, or they can use columnar storage. Teradata also provides a join index to speed up known joins. Even more incredible, they have intelligent memory where the system automatically tracks the most popular data that users are accessing that week, and they put this hot data in memory until it is no longer hot.
And Teradata has the QueryGrid! The QueryGrid allows the Teradata BYNET to connect directly to other systems, to create a logical data warehouse. Here is what Teradata has to say about the QueryGrid.
“The QueryGrid adds a single execution layer that orchestrates analyses across Teradata, Teradata’s Asterdata DBMS, Oracle, Hadoop, and, in the future, other databases and platforms. The analysis options include SQL queries, as well as graph, MapReduce, R-based analytics, and other applications.
Chris Twogood, Teradata’s vice president of product and services marketing, says, “The QueryGrid gives companies a transparent way to optimize the power of different technologies within a logical data warehouse.”
Teradata isn’t the only vendor building what Gartner calls the logical data warehouse. Recently, SAP introduced its Hana In-Memory Data Fabric for federated data access across sources. And since 2009, IBM has offered its DB2 Information Integrator for federated access to multiple data sources. And Amazon recently announced federated capability with its combination of Amazon Redshift, Amazon Aurora, and RDS for PostgreSQL.
It is as if every vendor is trying to integrate all systems to form one logical data warehouse.
So, if Teradata is so good, why has its stock gone from a high of 81 to today’s 26?
There are two major reasons Teradata is in data hell. They have the best overall system, combining top hardware design with great database architecture, but they have not understood the point-of-view of the desktop. Their query tools look like something out of the stone age. If you write the SQL for a Teradata system, it will return an answer set. Wow! And after 40 years in business, this is what you give to customers?
Teradata’s used to charge customers for its query tools (SQL Assistant and Teradata Studio), but when customers started switching to Nexus, and Teradata could no longer compete, they gave away its tools for free.
But the biggest reason for Teradata’s downfall has been the cloud. They were in a position to be the original creators of the cloud, but instead, it fell to Amazon, a company that sold books.
Teradata does now have a private cloud that they manage, and I highly recommend it for customers who want a world-class enterprise data warehouse.
But here is the cloud scoop. When someone rents a Teradata system on the Amazon, Azure, or the Google cloud, they are not able to use Teradata nodes or the Teradata BYNET. Teradata systems without the BYNET is like giving Superman kryptonite. The nodes and architecture are slow, and there is no QueryGrid.
Is a Teradata system on the Amazon, Azure, or Google cloud the best? Absolutely no way.
When it comes to creating a logical data warehouse Teradata, Amazon, IBM, and SAP all got it wrong. Steve Jobs once said, “Software beats hardware every time!”
My team at CoffingDW has spent 15 years building the Nexus Query Chameleon desktop so that everyone, from a basic business user to a top data scientist, can move data between any two systems with the click of a single button. Users can also drop-and-drag tables together from any system within their entire enterprise and Nexus joins the data automatically using over 100 different data movement techniques.
Nexus is the best software in the world, and that is the downfall of Teradata.
Why purchase a single system when you can join any data, at any time, from anywhere?
The only software solution that works is one that includes a brilliant desktop tool that allows users to do everything, but also allows the user to execute big data jobs from a server.
We now sell the Nexus Desktop and the NexusCore Server so a company can buy thousands of desktops and dozens of servers to move data between systems, perform federated queries, and run point-and-click federated analytics. The NexusCore Server is quite similar to a QueryGrid without the need to connect the hardware to the BYNET. The NexusCore Server is a QueryGrid for all technologies.
The next time someone tells you have difficult it is to migrate data to the cloud, or switch platforms, please tell them this is easy, and pass them my name and number.
How easy is it to perform federated queries? Drag in a Teradata table with an Amazon Redshift table. Then place a checkmark on the columns you want on the report and press Execute.
How easy is it to migrate your data from Teradata to Snowflake? Press the big green button (Execute) and find out.
If you want your users to access any data at any time, anywhere, then give me a call.
CEO, Coffing Data Warehousing
Direct: 513 300-0341
Youtube channel: CoffingDW
Tom Coffing, better known as Tera-Tom, is the founder of Coffing Data Warehousing where he has been CEO for the past 25 years. Tom has written over 75 books on all aspects of Teradata, Netezza, Yellowbrick, Snowflake, Redshift, Aurora, Vertica, SQL Server, and Greenplum. Tom has taught over 1,000 classes worldwide, and he is the designer of the Nexus Product Line.