 Hi, everybody. This is Jeff Kelly coming to you from Wikibon World Headquarters here in Malibur, Massachusetts. So when we think about big data, we often think about approaches like Hadoop and massively parallel analytic databases. But in a lot of cases, there are enterprises that are struggling with making a better use of their data in their more traditional databases. And not every enterprise is in a position now where they actually need these new big data technologies, but again, are really just trying to push performance and get some new insights out of the data they're already currently working with. So today, we're joined by Fred Sabokta, who is going to join us to talk a little bit about these issues. Fred is with FRS Consulting. He's both an IBM Gold Consultant and IBM Big Data Champion. Fred, welcome to the queue. Glad to be here. So as I mentioned at the start, so the big and big data tends to get most of the focus these days. But I understand you're working with clients that really aren't necessarily at that point where they're working with huge volumes of data, but nevertheless have some performance issues they're trying to work through and are always looking for new insights and data. So can you tell us a little bit about the types of challenges you're seeing in your client base and the things they're working with? Certainly. Certainly, what I'm finding in my customer base, first of all, my customers are a mix of small, medium businesses, Fortune 500 companies, law enforcement, public safety, criminal justice. They have traditional needs to be performed for their online users and also integrate well with business partners or other agencies or organizations. So there's been a predominant move toward increasing performance and getting the most out of their database licenses. Those are the most expensive licenses they purchase in their organization for software. So we want to make sure that they're getting their money's worth out of their existing database engines. So that tends to focus on performance tuning and also finding out how much we can get out of the data where it sits in the production transactional databases. And it turns out you can actually get quite a lot of insight if you are careful to tune your queries and get the information from the source. So we're finding ways to get data spread around the enterprise more quickly through distribution queries, asynchronous messaging, and also we're using a lot of XML-based features in the database to present more sophisticated, complete data sets in one pass to inbound and outbound data posts. Interesting. So it's, again, they're not really looking to these new types of technologies. They're looking to get more value out of what they've already invested in. So why don't we dig in a little bit more? You mentioned a couple of ways they're trying to do that at the end of your first answer. So why don't we dig into maybe some of the more popular ways you're seeing or the more effective ways you're helping your clients really get the most out of those database technologies they've invested in and are really critical to their business? Well, I'd say the absolute top priority is monitoring, not just for availability, but for performance. And those are different in terms of how they're done and also alerting on any kind of condition that could be great performance or create an outage. So for quite a while now, pretty much since the internet made systems 24-7 that weren't previously operating in that capacity, finding that a lot of shops just don't know what normal looks like. So we're working on instrumenting the key performance indicators of not just the database engine, but also outlying levels of the software stack and the business itself. You can sometimes find out things like if orders aren't coming in at the same rate, it could actually signify a problem with, say, a business partner or a supplier and you can take advantage of that situation and solve a problem, even if it's not originating from your side. So knowing what normal looks like is a huge part of knowing what you can push the system to do for the same license that you already paid for. If you have some fear or suspicion about, I don't know how hard a certain workload is going to hit this system, you're not going to be as willing to investigate leveraging your querying tools and your reporting environment and other data integration technologies if you're not sure if your system is performing consistently. So we tend to instrument that data very quickly with anything they have laying around. A lot of times we end up using monitoring software from the networking side because they tend to be the most vigilant about the performance and health of the different pieces of equipment on the network. And databases aren't generally monitored by those systems out of the box, but they can be very easily extended to monitor databases. And in that situation we end up with a very good picture of what the health and typical data of a database is like. Then we can figure out where are the opportunities to hit it with analytical workload that you wouldn't necessarily want to just do any time, but finding spots where you can extend the use of the database without stressing any of the existing workloads. So it's really the kind of that old saying knowing is half the battle. You really have to understand what's happening in your system. And then as you said kind of identify those, I guess what you call them opportunistic areas where you can take advantage of maybe some available resources to drive some of these additional additional jobs you want to do. But definitely the other side of it is also to make sure you're not over configured. You're unlikely to hear from your vendor that you may have licensed too many CPUs of a database or you've got too much going on with the server itself or you're using a virtualized environment. Can you step down the number of cores and still deliver a good response to all of the different workloads that it serves? You need to have those numbers, but also to see the trends. And so the best monitoring implementations I see where folks are getting the best response and use of their transactional environment is when they also have a record of how these things are being tracked. And I tell you it does not require high-end database licenses or very advanced performance tools to capture that. The networking guys once again have come in with software that helps you really develop a long-term understanding of your performance and utilization trends so you can plan accordingly. It's going to help with companies that have very sporadic budget cycles or long budget cycles. Part of the problem with big data and Hadoop and the other aspects of it is that they only can purchase technology once or twice a year perhaps. So there's a lot of planning that has to go in. And then when you arrive at that point and you have to set up your budget and make your commitments, you need to know what exactly your target is. So the data you get back from monitoring your workload and monitoring outages and monitoring other issues that happen in your production environment could really help you make educated decisions whether you're you're all ready to tackle a big data endeavor or if you're still in the traditional transactional space of maybe a data warehouse or a data mart. And you just want to make sure that you're using your resources efficiently. Yeah, interesting. So you're an IBM Gold Consultant, IBM Big Data, or I should say IBM Data Champion. So obviously you're very well acquainted with IBM's portfolio of data management products. You know, we've covered them, covered IBM quite heavily in terms of their big data strategy over the last six months to a year. I wonder what is your take on how IBM is actually helping their clients with their, I guess you might call it legacy, a data management portfolio, things like DP2. How are they, do you think IBM's doing a good job of kind of balancing the innovation they're doing on the big data side while still supporting kind of their customers who are using maybe not quite as sexy technology, something like DB2, that again are really core to their business. How would you grade IBM? What are they doing well? And what do they need to improve maybe? Well, first of all, I think it's interesting now that anything that's not big data is suddenly legacy because that's a neat take on things. I know there's a lot of hype around it, but to go as far as to say that if you're not running Hadoop, you're writing a buggy is possibly a bit premature. No, definitely I'd say as an IBM expert focusing my practice primarily on IBM of database technology, it's a very good spot. What I found consistently over the years is that IBM has done very good research and has done very thorough research in the improvements and enhancements that they introduced into the DB2 engine, both on the mini-frame side and on the Linux Unix Windows side. A lot of people don't realize that DB2 runs on Linux Unix Windows and it actually does very well. But it used to be that something would start off on the mainframe and then it would be adapted to DB2 on Linux Unix Windows and that's not so much the case anymore. There are different hardware technologies that they can exploit. Power processor line and their P-Series servers has features that DB2 can exploit very well at the hardware level. And also there are SIMD instructions in the Intel set that can work very well with certain types of database workloads. So IBM is putting the money and the research and the talent on developing, continuing to enhance the DB2 engine. It's just gotten incredible. I happened to be at the DB2 10.5 announcement earlier this month at the Almond and Lab. And I tell you it felt a little bit like James Bond getting the briefing from Q on all the new gadgets that are made available to him to do his job. It's pretty incredible what's been coming out of IBM for the past few years, especially the pure scale announcement that they made in 2009. Essentially a faithful, solid, mature implementation as parallel suspects for non-mainframe environments is a huge step forward. And in terms of delivering massive transactional volume in an active cluster is no small feat. And they really nailed it. And you see over the years IBM tends to not need to backpedal the direction that they choose for things. When it comes to the core DB2 engine, it's a very solid piece of technology. And I routinely bet my finger on it. I will bet a finger on DB2. Right, yeah, I was also at that event. And of course one of the major announcements was around blue acceleration. What was your take on that? As it applies to bringing, I think what IBM called speed of thought or speed of business analytic capabilities to DB2. Does that something that you think is going to have a lot of benefit for your clients? Sir, could you repeat the question one more time? I was wondering what your take was on blue acceleration announced at that Almond event, the idea of bringing more, I think what IBM calls speed of thought or speed of business analytics interactive queries to DB2. And I was wondering what your take was on that technology as it applies to DB2. And do you think that's going to have a significant value, hold significant value for your clients? So the DB2 blue announcement, which is essentially bolting on a column store to the relational and hierarchical XML part of the DB2 engine, is a fairly radical step forward. It really addresses a gap that was significant in the past and it was a gap where a lot of, not a lot of, but it was a gap where competing vendors with competing technologies were having an advantage. And the blue upgrades of the DB2 engine is just fantastic in what it allows by reducing the cost of the types of queries that used to be particularly extensive to run in your prime transactional database. Sorry, go ahead. You want to answer that again with the video? No, no, just please just continue. So in terms of, actually, do you foresee that as something blue for DB2? Actually something you're going to see your clients start adopting or is this a little bit more maybe kind of future road map kind of thing for your clients? I think as soon as clients understand, the DB2 users understand these scenarios and music patterns where the blue technology offers a serious advantage, then you'll see them adopting it. Because like the other parts of DB2, that when the engine has been upgraded to do new things, it's been done at such a core level that you can mix and match different storage technologies inside of the DB2 engine, have a single query going off to XML, also to multi-dimensional clustered row tables in one area, and then now with blue tables in the other. These are first class citizens in the DB2 engine. So when you have a workload that could benefit from structuring one table, organizing it by column, and not having to trade off with the other tables that aren't necessarily designed to benefit from that, it's going to be a much lower barrier to entry for organizations that are using DB2 and are confident with it and they understand what it can do for them. So we've got time for just one more question. And I wonder if you could just give one piece of advice to CIOs who are kind of struggling with coming up against performance issues in their data management infrastructure. Like some of your clients, they're looking for new insights but have limited budgets and maybe aren't in the position where they're going to start investing in some of these big data technologies like Hadoop. What one piece of advice would you give to those folks who are looking for help there? Well, if you've got a vendor database and you have not mastered it, you've not confident in understanding every aspect of it and what it can do for you for the same price, if you're not making total use of your database and if it's because your staff aren't versed enough on that database engine to make use of those. I mean, if you have strategic issues or technology issues that are preventing you from fully exploiting the product you pay for, that's one thing. But if your staff are not up to date on the latest innovations in that database engine, then you're really cutting yourself off at the need. So I'd say that a fairly affordable way to get more performance out of your software and hardware is to make sure that your IT administrators are trained enough to date on the latest features. That may not necessarily be classroom training. In fact, most of my clients lately have gotten the most benefit from attending technical conferences. And in the case of DD2, if the iDoug conference, which is happening next week in Orlando, there's also an IBM conference that centers heavily on DD2 and other IBM products. If you get your staff trained and current on the technologies, they're going to take better care of the system. They're going to exploit the features that you're already paying for in your database. And it could mean compression cutting your IO. It could mean compression cutting your storage costs or a blue table that suddenly makes a very expensive, long-ranging scan a lot more affordable to run in production at any time. So definitely make sure that your people are armed with all the knowledge they need to leverage the tools they already have. Okay, some good advice. Fred Saboka, thank you so much for joining us. Fred from FRS Consulting talking to us today about some of the technologies and some of the approaches you can take to kind of get the most performance and find some additional insights in your database environment. So Fred, thanks again for joining us and thanks everybody for watching.