Ron Powell, independent analyst and expert with the BeyeNETWORK and the Business Analytics Collaborative, interviews Alan Greenspan, product marketing manager at Teradata. They discuss why the best database is now also the most flexible database.
Alan, I recently saw the announcement about Teradata Everywhere. What is Teradata Everywhere?
Alan Greenspan: The Teradata Database, to our knowledge, is the first database to be available across a variety of specific platform types, which include the public cloud, managed cloud, private cloud with VMWare as well as workload specific appliances on premises. This gives our customers a lot of flexibility because they are now able to put the same database in all of those deployment and associated business models. So our customers can now deploy the architecture they want, the business models they want, whether theyíre purchasing through the cloud or on-premises hardware, and mix and match them. And, they can consider things like data gravity. If they have a source system running in the cloud creating data and they want to put the analytics closer to minimize data movement overhead, they can put an instance of Teradata wherever they need it. We now have all the systems and all the deployment choices that a customer might want.†
Everyone is familiar with Teradata on premises. Letís get into some of the specifics regarding the cloud, whether itís a managed cloud, a private cloud or the public cloud like AWS. Could you get into the specifics of the announcement?
Alan Greenspan: The announcement is that we have the Teradata Database that people have been using with all the capabilities, the very mature optimizer, and same feature set Ė it really is the same code base in all of these environments. It gives the customers application consistency. It is not just a branding exercise across separate platforms and separate database products, which some vendors do to make it look like they have a product family. This really is the same database.
What we did is develop very native, easy-to deploy configuration software, rather than change the database. Weíve done a lot of engineering work to deploy the product, make it run natively in each of these environments. Itís the same database product that youíre getting that is using the same optimizer that is massively parallel with the same functions and advanced features, and ease of use and administration that people have been used to with Teradata.
So regardless of platform, it is the same MPP database on all.
Alan Greenspan: Yes, it really is the same code base, from the same engineering organization. So as we add features, those features will release on all of the platforms.
Does that make it very easy then to migrate from one platform to the next Ė to move from on premises to the cloud?
Alan Greenspan: Thatís one of the reasons we call it Teradata Everywhere. Itís as if it is everywhere within your environment, and if you decide to move Ė think of it as future-proofing the environment. If your business grows or changes and now it makes sense to move it to the cloud or to expand or add another system to the environment, it is the same. Itís almost as if it was floating everywhere. It doesnít matter from an application point of view where you point the query. Itís more of a data and environment architecture.
How significant is the Teradata MAPS architecture enhancement in the cloud with VMWare?
Alan Greenspan: It is significant. So within the Teradata Everywhere announcement, we also announced several supporting features. The Teradata MAPS architecture lets us expand systems seamlessly with the system online. Thatís great for on prem. It takes away almost all of the offline time in expanding a system. But in the cloud, it also gives us elasticity so we can now move data from node to node. What that lets you do in the cloud is if you have a cloud burst situation where you have a peak load Ė either for a quarter end or a special analytical project Ė you can expand the cloud system. You add more nodes to it, and move data to it. Now when you get past your quarter-end processing or finish the special project and that peak load subsides, you can move the data back to a smaller set of nodes, release the footprint and stop paying for it. So it lets you expand or shrink the actual database. When some databases are put in the cloud, they canít expand or shrink in the cloud. So you need to rent a bigger database and copy your whole database into that new bigger instance. Weíre not doing that. Weíre treating it like we do on prem. With a massively parallel system, we can expand it. In the cloud, with the MAPS architecture enhancement to the Teradata database, you can expand more nodes and you can also push the data back to fewer nodes and then release it.†
So it gives our customers great flexibility, but the queries donít change. So the database internally knows how to access the tables and the data when you submit your query. A query that worked when the database was 12 nodes now can work in 16 nodes. If you go back to 12 nodes, you havenít touched your SQL at all. Itís very transparent.
Alan, youíve been with Teradata for quite a while. Is this a big culture shift for Teradata to be on all these platforms?
Alan Greenspan: It is, in a sense. Weíve always been very customer focused. And, the marketplace and our customers are moving to the cloud. Having the flexibility of the cloud is great for business. From a customer focus, itís natural that we go there. Having these various platforms and the mix of platforms is somewhat of a culture change, but itís happening very quickly. Fortunately, the organization has moved aggressively. We released Teradata on AWS in the first quarter of 2016 as just an SMP single-node version. In September of 2016, weíve announced MPP on AWS up to 32 nodes, true MPP on Microsoft Azure, and true MPP on VMWare for a private cloud up to 32 nodes. So very quickly weíve gone to having very large production systems in these environments. Weíre back to that kind of speed, and itís great to see. It will give our customers a lot of headroom for the future.
Well, I would imagine that having the same database on all of these platforms gives you the agility to keep moving quickly.
Alan Greenspan: It does, and I think there are some people that are going to be surprised that the best database is now also the most flexible.
Alan, thank you so much for giving us an update on Teradata Everywhere.
Recent articles by Ron Powell