IBM earlier this month unveiled its first update of DB2 software in four years. With the new database software, IBM is on a mission to tame the data deluge.
We caught up with Bernie Spang, director of strategy, and database software and systems to discuss the latest iteration of DB2, how it's attracting customers from Oracle and some practical examples of DB2 10 in action.
Q: How much success are you having getting database clients to migrate away from Oracle?
A: We're seeing success continue to accelerate quarter after quarter. Clients are looking for alternatives. They have been, in fact, for some time but many have not seriously considered moving because of the perception that it's complicated, risky, costly, etc.
With the compatibility for Oracle database applications that we introduced in the previous version of DB2, we really changed that equation. With DB2 9 we also upped the game for optimization for SAP application workloads. We're starting to see that snowball effect really taking hold.
Q: With DB2 10, you stressed a few specific benefits. One of them was accessing Big Data for deeper insights. How deep are we going with this? What's different or new in DB2 10?
A: Hadoop-based systems use a different approach, a file-based distributive approach for analyzing structured and non-structured information. And that's important. In the last few years we have entered into a new era of data management where the answer isn't automatically "get me a relational data system." It's "get me the right system that best handles and can analyze this particular type of data."
In this Big Data era where we are dealing with the high-volume, velocity, and variety of data we don't want to create silos of analyses and offer different results in different pieces back to the business user to put together. We don't want to recreate the problem that was caused in the early days of business computing where each different application has its own database, a known data structure, and that has to be integrated. That becomes a big project.
In the era of Big Data analytics, we want to be smarter than that. We want to think ahead. We want to model all the insights from these different systems to be able to easily flow together. By combining them you get even greater insights, new insights. New enhanced integration enables you to come from within the warehouse to access the insights that you're generating from Hadoop-based systems, bring those to the warehouse and add them to the mix of your analyses of the structured data. Then you can feed back insights to the business user, a more insightful answer.
Q: Let's talk about how DB2 10 speeds up processes and lowers data cost. People love to hear faster, lower cost. How are you accomplishing that?
A: First of all, we're extending our industry leading compression capabilities with what we're calling adaptive compression. Early access clients and partners saw upward of 90 percent of their freed up by compressing data. That obviously offers cost savings. Remember, storage is not just the cost of the disk but the space, power, cooling, administration, and everything around it.
Beyond cost savings, it actually delivers better performance because you're moving less data in and out of storage. That has a positive effect on your performance. We've also added a lot of new capabilities within the system itself focused on accelerating query performance.
And we offer continuous data ingest capability with this release. That's using parallel processing to support a continuous feed of data. This is particularly important in the warehouse scenario, so that you are able to continuously feed data into it without impacting the performance of the analytics query going on in the warehouse.
Q: What about decision-making? How does the new iteration help in that regard?
A: Our new time travel query capability lets you access and analyze information at different points in time -- both past and future -- in addition to the present. This dramatically reduces the amount of application code that needs to be written for this kind of time-dependent application.
If you're reviewing an insurance claim, for example, you need to understand the terms of the policy that were in effect at the time of the accident, which may be different than they are now. You need to be able to access that information at a point in time in the past.
If you're in the travel business and you need to review travel itineraries, which have future dates, you may find some issues. For example, if you've got a hotel booked for a week in Rome and you have a car service booked in New York City next month at the same time, then there's something wrong with that. You can quickly find out and fix it.
If you're in retail or manufacturing, your product pricing or the price of goods that are going into the manufacturing process are going to change next month. What is the cost? What is my profitability going to look like next month versus what it looks like now? We can help you get those answers.