 Welcome to Newsdesk. I'm SiliconANGLE TV. For Wednesday, October 10th, 2012, I'm Kristen Folletti. Watch out, Larry Ellison. IBM is taking on Oracle with peer data appliances. What does IBM have in store for its nemesis? Here with his breaking analysis is Chief Wikibon Analyst, Dave Vellante. Welcome, Dave. Morning. In a move to compete with Larry Ellison and Oracle's Exadata appliance, IBM is expanding its line with a family of boxes called peer data. Can you give us some background on this new product line? Sure. Last year, IBM announced its pure systems. Actually, it was earlier this year. And this extends that line. It's basically IBM's play on converged infrastructure. IBM announced a general purpose version of the platform really to compete with VCE and Vblock and HP and NetApp. And also announced something called pure applications. And the implication was that they would be able to tune these systems and bring application specificity to these boxes. And that's really what the announcements yesterday were all about, the pure data, really bringing both transactional and analytics applications to the converged infrastructure play. Some IBM executives have joked in the press that this is the Exadata killer. It's interesting to note that the announcement's the week after Oracle open world. I will tell you, Kristen, that the similarities are greater than the differences between these two companies. Oracle has the red stack. IBM has the blue stack. Oracle ships its products with the Oracle database. IBM ships its products with the IBM databases, DV2 and the TZ databases, et cetera. IBM is somewhat more open than Oracle. It allows other people storage to attach to some of these systems. But in general, both companies are trying to essentially make the value proposition so attractive and compelling to users that they will go nowhere else. That's a euphemism for lock-in. And that's really the key customer consideration here. But the systems are appealing, both the Exadata and the IBM Pure Systems. They're appealing and alluring from the standpoint of they take away a lot of the mundane integration tasks that the customer has to do. And they come pre-packaged and pre-integrated, which makes them easier to install. And of course, IBM and others emphasize the time to value, the time to deploy applications is much, much faster. Oracle's co-founder and CEO, Larry Ellison, has said more than once that he wants Oracle to be like the IBM of the mainframe era. Can you tell us about IBM's strategy with Pure Systems and how it compares with Oracle's? Yeah, so of course, IBM and its heyday, Kristen, had about 50% of the industry's revenue and about 2 thirds of the industry's profits. Even Microsoft's monopoly was not that powerful from a relative standpoint. So basically, what Larry is saying is he essentially wants to have that market power the way that IBM had that market power. Really, the differences between these two companies are that IBM is really not approaching it as a one-size-fits-all. It really emphasizes horses for courses. Now, Oracle as well has its version of horses for courses. Horses of courses is of course a horse racing analogy where some race sources do well on certain race courses that maybe, for instance, have tight turns and others do well on ones that are wider turns. And so what IBM is stressing is that it's designing different products for different applications. Now, they'll attack Oracle in the sense that Oracle uses the same database essentially throughout its line, but Oracle does configure and tune its systems for different applications. It just uses the same basic database infrastructure whereas IBM will use DB2, for example, for transactional systems. It will use Netiza and an acquisition that it made in 2010 for other data warehousing applications. Essentially, what IBM announced yesterday is our systems that are transactional oriented, that's the DB2 version, and more data warehouse and analytics oriented in two versions of that system. One that can do thousands of transactions per second and queries per second rather, another one that can do many, many complex queries in seconds. So essentially three different product lines tuned for different applications really trying to attack Oracle's one size fits all approach. Where does IBM stand in the whole converged infrastructure play? So converged infrastructure is this new trend that we've been tacking since about 2009. Essentially what it does is preintegrate storage and servers and networking into a package, into a single block of infrastructure. Oracle was one of the early ones, certainly Netiza was there early and VCE was really the first to come out with a general purpose version of that type of system. And where IBM fits is it's playing both head-to-head competing with VCE with expert integrated systems, a general purpose system that combines storage, networking and servers, and it's competing with Oracle in the application business, both for transactional applications with DB2, its mainframe and now open systems database for transactions, as well as analytic and data warehousing applications. So IBM is playing in the entire stack from hardware all the way through applications, whereas companies like VCE and NetApp and to a large extent HP, not withstanding Vertica, are really competing for the infrastructure play. What does the competition between these two lines mean for customers? I think the choice for customers is not straightforward. On the one hand, you've got the allure of being able to take pre-packaged and integrated technology and drop it in and you don't have to do the integration yourself and that is very appealing. On the other hand, it is the mother of all lock-ins. Now, the other real, probably more important consideration for customers in our opinion at Wikibon is that the general purpose approach is more fungible. In other words, you can support more applications across the portfolio and even a role your own approach. For instance, buying an Oracle database, running it on an HP server and buying EMC or NetApp storage, best to breed the equipment that you can integrate on your own or with the help of a service provider. That type of infrastructure is going to allow you to support more applications across the infrastructure and ultimately could lead to a lower TCO over a longer period of time because you're going to get better return on assets because that infrastructure is more flexible. At the same time, if you really want to do one job well, these appliances are really catching on. I mean, think of knives, right? It's the Swiss Army knife versus the bread knife or the butter knife or the steak knife and that's really the trade-off and choices that customers have to make. What are the real goals for these peer data appliances? What capabilities do they need to have to truly stand out from their competitors? I think the real differentiation that IBM has that it emphasizes, it's bringing applications, expertise from its applications business and its services business and embedding that knowledge into the system. IBM markets smarter analytics, the smarter planet and it's trying to differentiate on that knowledge, trying to codify that knowledge into the system. Now it's the early days, but IBM does have templates that it loads into the system such that the system has the intelligence to configure itself and actually operate on the data in a way that is most effective. So that's IBM's primary differentiator. As well, another benchmark that we're watching is how well is IBM able to hold on to those existing, for instance, DV2 franchise and systems franchise. So this is both a defensive play to try to protect a large install base and an offensive play to really try to go after Oracle. How does Hadoop play into all of this? Well, we only heard a little bit of talk about Hadoop yesterday at the IBM announcement. Generally speaking, I would categorize it as follows. The traditional legacy companies, the database companies, the data warehousing companies would like us to believe that Hadoop is batch and they are really the real-time component of that. Now, IBM again is a little less dogmatic in that philosophy, but I think in general falls into that camp where it would like you to use Hadoop as the storage processing batch engine and then use its systems and its analytics and its Cognos and its data warehousing technologies to create real-time. Certainly Larry Ellison's demo at Oracle OpenWorld is of that ill. And so I think the fact of the matter is that Hadoop is growing, it's growing, probably orders of magnitude faster than the traditional database and traditional data warehouse business. And it's a reality that all companies are going to have to contend with. The philosophy or the approach that most have taken is to build connectors into Hadoop so that you can connect traditional legacy systems into Hadoop, pull the data out and do the processing through an ETL or a transform and load operation. There are others, for instance, like Hadaap that is coming to market and you'll hear more about this over the coming months, that is trying to bring both the traditional SQL database world and the Hadoop unstructured world together to a unified system. I've seen others talk about that as well, such as Otivio. So there are some emerging startups that are really trying to solve that problem and really trying to make real-time Hadoop a native capability. Well, Dave, we appreciate your time today. Thanks so much for joining us. My pleasure, thanks for having me, Kristin. Keep up to date with the latest in tech innovation by joining us daily at Newsdesk on SiliconANGLE.tv.