Skip to content

Big Data industry titan, Teradata, pursues a truce amidst fierce industry conflicts, commonly referred to as the "Holy Wars."

Data management software arena, once tranquil, has transformed into a contentious battleground due to ideological divisions. On one end, there are the traditionalists, who staunchly support their methods and values.

In the ongoing battles of big data, Teradata looks for a resolution
In the ongoing battles of big data, Teradata looks for a resolution

Big Data industry titan, Teradata, pursues a truce amidst fierce industry conflicts, commonly referred to as the "Holy Wars."

In the ever-evolving world of data management, a fascinating debate is unfolding between traditionalists and upstarts. This conflict, which was evident at a recent conference, revolves around the question of whether relational databases or non-relational platforms like MapReduce and Hadoop are the future of enterprise data.

Stepping into this fray is Stephen Brobst, the chief technology officer of data warehouse vendor Teradata. Rather than taking an ideological stance, Brobst advocates for an enlightened view, believing that a good engineer should be open to using good ideas from any source.

One such idea that Brobst finds particularly useful is the concept of 'late binding', a feature from the MapReduce world. In this approach, data is extracted from source systems in its raw form, and the structure is only imposed when a query is sent.

This method is exemplified by eBay, a company that processes vast amounts of data daily. Traditional ETL programs would struggle to keep up with eBay's data tracking needs. Instead, eBay uses MapReduce for web analytics, a decision that has proven to be beneficial.

eBay tracks over 10,000 tags in their weblog data and wants to track new things every day. The 'late binding' feature in MapReduce allows eBay's analysts to apply structure to data at query execution, providing the flexibility needed to adapt to the fluid nature of data structures.

In contrast to this, Teradata views Hadoop as a low-cost repository for storing all data forever. However, Teradata doesn't abandon its relational database roots. Instead, it incorporates MapReduce technology and the 'late binding' feature into its product offerings, as demonstrated by its acquisition of big data analytics provider Aster Data in 2010.

Aster Data extended the SQL model with MapReduce and the 'late binding' feature, without discarding the productivity and tools offered by relational databases. This approach allows Teradata to cater to both traditionalists and upstarts, offering a solution that is adaptable to the rapidly changing landscape of big data.

In summary, the data management software industry is witnessing an open battle between traditionalists and upstarts. However, innovative thinkers like Brobst are advocating for a more pragmatic approach, one that values flexibility and the ability to adapt to the ever-changing nature of big data.

Read also: