I am attending the Teradata Influencers Summit in lovely Del Mar, California. First up is Oliver Ratzesberger, Senior VP of Software, to talk about recent technology innovations and technology strategy. Oliver highlighted some of the key trends and themes for Teradata. He began with a maturity model for “the sentiment enterprise” (video here):
- Agile Data Platform
Managed but agile data.
- Behavior Data Platform
Adding behavioral data to static data
- Collaborative Ideation Platform
Making it easier for people to collaborate on data
- Analytical Application Platform
Make it easy to make analytical applications
- Autonomous Decisioning Platform
Decision Management really. I am not sure this has to be level 5 – plenty of folks out there building powerful decisioning systems who haven’t built this kind of agile analytical platform first.
The technology Teradata is developing for this vision is driven, he says, by a set of core tenets:
- Cloud-first and cloud design principles
- Self-service, open APIs and common ingest/consume layers
- Open source leverage and contribution
- Simplifying infrastructure e.g. through QueryGrid
Increasing the agility of the Teradata platform, making it easier to expand usage and evolve / adapt it, has been a core focus for the last few years. Datalabs, support for new analytic languages, cloud and VM support, REST APIs and JSON etc are all about making the platform more integrable – not really agile in my mind but making it more ecosystem-friendly. Making it easier to integrate with lots of data sources or types, providing more real-time options, making it easier to integrate into analytical applications etc.
The Teradata UDA has been focused on integrating the data warehouse, hadoop and the discovery platform. Increasingly this is being extended to also focus on real-time processing in addition. The architecture needs he says to support different kinds of data—from tightly coupled, strongly managed data to loosely or even non-coupled data. The data infrastructure needs to support all these kinds of data so that organizations can bring in data quickly (non or loosely coupled) and gradually increase the degree of coupling or integration. This journey needs to be easy and the UDA is designed to support this along with QueryGrid etc.
Increasingly this architecture is being delivered through analytic apps and Teradata is investing in making it easier to develop apps with the Aster App Center and other plans. Part of making it easier to develop apps is removing the need for ETL processes. Teradata is working on Teradata Listener, a governed self-service platform for collecting data from applications and pushing it into the UDA data stores.
Lots more to come apparently.