hello and welcome thank you for attending our customer spotlight webinar how well care accelerated big data delivery to improve analytics I’m happy to introduce our presenters for today David Menninger from pivotal Kevin Petrie from attunity and james clark from wellcare please feel free to send in your questions any time using the QA console you’ll do our best to answer them at the end or if we can’t get to them today we’ll get back to you via email thanks again for being here and now I’ll hand it off to David thank you today we would like to review with you some of the challenges in big data and the opportunities not only for healthcare providers but in general help you understand how to take advantage of big data and in the case of healthcare analects specifically how to increase some of your reporting speeds with all the compliance reporting needed in healthcare being able to prepare reports quickly and easily is important enabling real-time data initiatives for big data taking advantage of the data that is streaming through your system and being able to react to that in real time incorporating Hadoop into your enterprise data warehouse environment and how you might utilize the dupe to achieve something referred to as a data lake and might do that successfully and talk about and with the help of our friends from old care talk about how on a joint solution leveraging both the attunity and pivotal big data capabilities when you’re setting up these in so the first thing I’d like to do is introduce pivotal give you a sense of who we are not everyone I mean was I meet with hundreds of customers in my role and not everyone knows who pivotal is they may have heard of pivotal as being a start-up and want to dispel the notion that we’re a small start-up we were founded two years ago put together with parts of EMC and parts of VMware we have over 1,800 employees now part of the organization and over thousand enterprise customers so this is not three guys on a dog in a garage they well-established analogies they lie on in your is and there’s some revenue figures cited their sense of how big the organization is from a licensing perspective the problem that we’re trying to address as an organization pizzle is in fact the name pivotal comes from this transformation that we’re in the midst of as an industry every business has to become a digital business these days to remain competitive we all do our banking you know on our phones or over the web in the case of health care I go to a relatively small medical practice 3dr medical practice and even at that practice i can get my medical records online and over over the phone so this this world is changing and for you to be competitive in your business you have to be able to take advantage of that and it’s this pivot or this transformation that the name pivotal represents so pivoting to become a digital business and as an organization we saw the opportunity to help businesses across all different kinds of industries have that platform that would enable them to take advantage of it become in this world of big data as we talk more about big data will talk of data technologies typically require scale-out implementations meaning any servers operating together as a cluster of computing our living those capabilities to your users now companies like Google and Facebook and Yahoo pioneered these types of approaches they could afford to invest billions of dollars in building out server farms experimenting with open source technologies at creating some of those open source technologies most of our enterprise customers don’t really have that luxury well to experiment and explore what they might do and so we have created a platform that effectively gives you google like or Amazon like abilities in your enterprise so you can take advantage of being a digital and when we look at what is required in this type of platform you’ll see will will end up focusing most of our discussion around the Big Data part of the equation so that’s an obvious part of today’s discussion in other discussions we will talk we could talk to you about agile capabilities that you need to live in this world just take a look at your app store right now see how many apps need to be updated in your out of store now

to give you a sense of how quickly things need to change and how quickly you need to be raq be able to react to changes in the moment and as I mentioned scale-out technologies are often delivered in public cloud platforms but those could be private cloud platforms as well so you need a way to be able to deal with these elastic architectures and how do you provide services to your users in an elastic environment and do that in a way that is manageable and repeatable as an organization those are the things that we see constituting the key pillars if you will of this digital transformation and as I said today’s discussion is going to focus mostly on the big data portion of the equation the fundamental problem for most organizations in dealing with big data is that we have a lot of data as an industry that we just don’t utilize you can see here the statistics on the right-hand side we really only prepare three percent of our data for analysis we really only analyze a half a percent of that and even less than that is being operationalized by operationalized we mean that analysis is not a one-time exercise when you identify a way to analyze information and get useful value out of it need to be able to institutionalize that and make that a regular part of your standard operating procedure so it needs to be bedded into the business processes and the applications of your organization without that you’re left with just a one-off analysis and rapidly absolute an out of date with your analysis so you don’t keep it up to date and make it part of your standard on so the fundamental problem is how do we provide architecture and how do we provide analyses that can incorporate all this information and do it in a way that doesn’t leave data falling off the table and being discarded the pivotal has what we refer to as a big data sweet you see various components of the Big Data sweet represented on the screen here let’s concentrate on the three boxes across the top first the process of collecting the data often requires a bunch of manipulation and preparation of the data what we’ve called data engineering this data processing box we’ve got several components based around to dupe they have heard if you like many organizations I meet with about ninety percent of the organization’s I meet with the totally non-scientific number but about ninety percent of the organization’s I meet with are not very far along in their Hadoop journey yet there are certainly some exceptions you know digital media companies and Wall Street firms certainly are further down the path what to do but it’s not uncommon if you are listening to this broadcast and your concern that you haven’t yet embraced to do it’s appropriate to be concerned but don’t feel like you’ve missed the opportunity most organizations are beginning this journey down the Hadoop path and who do provide several things that are harder to do in more traditional databases a dupe is very good at dealing with unstructured data in the case of healthcare data think of images and doctors notes like that I dupe is also very good at dealing with data at very large volume and so those are the types of things that had died and as you work with a dupe it is a programming environment and having some tools like spring XD and spark available to be processing that data as you’re preparing it for analysis and as it’s streaming into the system those are valuable component but think of the left-hand boxes primarily be represented by Hadoop we also have in the middle box sequel based parallel sequel based advanced analytic capable so while Hadoop is very powerful it’s also much more difficult to use been sequel and so you probably have within your organization already a number of people with lots of sequel based skills so we offer those sequel based skills in two forms or sequel base capabilities in two forms the green plum database and Hawk Hawk stands for Hadoop with query so those same Greenplum sequel analytics are available running standalone in the green thumb database or as part of pivotal hawk running on top of to do and to give you a sense of the breadth and depth of the analytics that can be performed in mock and in the database we are one of only two database vendors that are certified to run SAS SAS SAS software analytics in the database so gives you a sense of the breadth of analytics that we can perform there similarly if you’re familiar with

the transaction processing council they have a decision support benchmark TP CDs the hawk greenplum database can run all 111 of those queries to find are not supported in most other tools the last piece of the puzzle on the right hand side is deploying those applications that said it’s important to operationalize the analytics that you create and so we have tools to create those applications at scale the pivotal gemfire component represented in the upper left part of the apps at scale box is an in-memory database if you think about Big Data part of the reason it’s big is that it’s occurring constantly so if you can react to that information as it’s happening that often spot opportunities and battalion healthcare world even save lives by looking at data in real time now several other components there the Redis and RabbitMQ components like spark and spring XD components you might use building out some custom applications or capabilities when you go to operationalize the types of analytics and the pivotal Cloud Foundry icon represented in there is our platform for deploying the applications so that may be a discussion for another day we offer all of these services across the bottom here you see they’re all also offered as components running within the gun foundry part of our portfolio couple of things to note about the Big Data Platform what makes it different from others I mentioned the sequel leadership on Hadoop already this is a single license across all the different components one of the things we’ve observed from our customers is that they don’t necessarily know which parts of the big data stack they’re going to need or use right away they might start with sequel then they might add Hadoop and they might add in memory capabilities or they might start at one of the other points and so this is a complete platform all incorporated into a single licensing mechanism so you can pick and choose which pieces you want to utilize they’ve got the opportunity to deploy those different pieces either stand alone or in this configuration with a dupe and running on top of cloud foundry these are all open source components or in the process of being open sourced and I’ve already mentioned the ends data to get so those are some of the differentiators of this platform and why organizations like well chair wellcare have chosen to work with this technology so with that I’m going to turn it over to Kevin Petri like Kevin patrians to to the part of the solution Kevin great thank you David so everyone I’m very pleased to have the opportunity to speak with you today for a few minutes about what we see happening working with healthcare and a number of other healthcare providers and health care related organizations there’s some fascinating things going on in this part of the of the Big Data industry and we’re very pleased to to share with you some of the things that we’ve learned so too the table briefly we are a publicly traded organization we trade on the Nasdaq on the do the ticker symbol ATT you we have global operations we have two thousand or more customers in 60 countries and we’re very pleased to be serving over one half of the fortune 100 we’re also pleased and it’s a responsibility going forward to be recognized by various industry experts for our innovation that’s something we fully intend to to continue to push the envelope with what we’re fundamentally seeking to do is help organizations manage big data more efficiently in order to free up resources to focus on data science so if we train the lens here on health care there are three key challenges and three key opportunities that we have seen working with various clients the first is that as we all know risk is moving due to legislation due to market forces from the patient to the provider that creates new pressure on deprived riders of all types to be more accountable and more focused on patient outcomes and the quality of care there’s this rising in tandem rising public expectation that it’s not acceptable to have patient care suffer in any fashion if the right data point doesn’t reach the right provider or the right doctor at the right point in time so there’s a pressure here there’s the converse opportunity as well that we’ll talk about the final challenge is really getting down to platforms like any organization like any enterprise health care providers have more than one platform the electronic medical records

movement is gaining stride is really starting to digitize records and put them into usable form as David pointed out the challenge remains to continue that and also to integrate those digits across databases across Hadoop in the cloud and so forth if we look at the opportunity side of the ledger there’s some pretty compelling things going on here it’s been and exciting to see the level innovation looking at smartphones looking at home based technology to create a very rich data stream from the patient wherever they are potentially in their home potentially going about their daily lives back to caregivers so the caregivers have an unprecedented opportunity to improve care both when the patient is within the clinic walls within the hospital walls and when the patient is out living their lives a second great opportunity here is that there are methods of improving operations which is critical for healthcare organization as for any other that have been proven in other industries so some of the proven methods include basic logistics if you will we have a client that we’ve worked with to create or we’ve they are helping their clients create the emergency room of the future essentially by treating a hospital like a factory floor and putting RFID tags on various individuals doctors nurses and so forth in order to monitor the flow of equipment and monitor the flow of digits and optimize it with future design that’s the type of approach that has worked in various industrial industries for a long time it can work in healthcare final point here and this is something we’re pleased to be contributing to as well is that the methods of integrating data continue to improve so let’s go to the mouths here this is some very interesting survey data from a burr green Aberdeen group earlier this year of healthcare professionals and they see a few key points of pain the first is that many critical decisions could benefit from more data-driven support no mystery there no secret or surprise another is that a lack of operational visibility for example into how data is being used does create some inefficiencies another key point here is that too many disparate data sources and data silos persist so again breaking down those silos is critical in healthcare as in other industries and finally the volume and the complexity of proliferating endpoints proliferating data types and proliferating platforms does create opportunities we talked about but it can also create some management complexity so the bottom line is that to extract advantage from big data you need to move it in order to gain that advantage it’s really a move it or lose it value proposition here Hadoop and the cloud are great platforms Dave talked about plat about Hadoop earlier Hadoop usually is not the starting point for data it’s where data goes after it’s generated elsewhere so some of the end points that generate actual data you might have transactional systems you might have point-of-sale systems social media streams and smartphones those the data needs to move move from there to a place where it can be analyzed and the method of doing that needs to become increasingly efficient what we at attunity propose is improving big data management on three specific dimensions in each of them feeds into the next the first is profiling usage in order to optimize placement we provide visibility software that can help healthcare organizations and other enterprises understand how data is being used within data warehouses so that they can profile that usage they could for example identify hot data or data and thereby move it using a replicate software to the right location based on that information that feeds into the next point which is that we can more easily integrate data more rapidly across multiple platforms once that data is in place it needs to be prepared for analytics and that where that’s where we provide software that can automate aspects of data warehousing we exist as a Switzerland if you will we support 35 endpoints in terms of sources and targets and that cross-platform capability moving from where a data starts or where it resides to where it needs to be in order support analytics is our fundamental value proposition we’re very pleased to be partners tight partners with pivotal we’ve worked

together on many enterprise accounts in order to feed data from different starting points in two pivotal HD into pivotal greenplum and thereby support application usage using just some of the pieces of software here have a bit of a pause on the interface here we can feed data into the pivotal database big data sweet through two primary methods using our replicate software we can do full loads of data and we can do change data capture so what that means is that PhD and Hawk can more easily receive information because we automate the process by which the pivotal target is reconciled with the source whatever it is we’ve removed the manual coding required to do that we can load the data very easily and then send continuous updates to support more real-time applications using change data capture will do this all through the pivotal parallel file distribution program so at this point I’m going to hand over to James James can tell us a little bit about what he has been doing at wellcare good afternoon everyone my name is James Clark I’m 90 director for well care to talk a little bit a welker itself we are a leading provider of managed care services really targeted for government sponsored programs like Medicare and Medicaid we are a fortune 500 company our membership is about 2.8 million members as of the end of 2013 our corporate offices are in Tampa Florida we have a little over a little over 5,000 employees we as a company we’re using some large and established technologies Oracle sequel server my sequel but we really had some need to enable big data analytics across large projects involving clinical member information claims lots of demographic information operational data for decision support for pricing controls and those kind of things one particular area where we saw a huge amount of results was in our quality reporting systems where queries multiple sets of processing queries were taking up to 30 days to run and we’ve seen that across several kind of verticals for us and due to the large amounts of data both sitting in the traditional Oracle sequel server we saw continued no servers processors storage memory requirements to continue to grow and have for several years and we are also in the initial stages of our Hadoop implementation were well down the path with having implemented pivotal greenplum Oh next slide here uh and really some of our goals were to improve the analytical capabilities and in particular to reduce that 30-day processing of queries that I mentioned before as well as to eliminate various CTL systems and ways of manipulating data to get them into data warehouses to be able to have the capability to analyze that data at a high volume and we needed a data solution that could move data from transactional systems into our pivotal Greenplum database quickly and efficiently with change data capture as well as full dataset loads what we were able to do in implementing both attunity replicate and pivotal Greenplum database was to show a dramatic increase in our capability to get data into systems and provide analytics on those I think the average of increase of reporting speed that we’ve seen across all of the processes that we’ve loaded into the system is about seventy-three percent we’re able to do roughly I think it’s about seven days of processing where you know before we were achieving that 30 days of processing as well as to reduce the complexity of the custom PL sequel etl

jobs and other various components arm and things like that of being able to pull in data to the various systems and you know we’re able to meet our regulatory requirements quickly to be able to analyze that data to be able to actually run multiple analyses of the data validate our results cut out some third-party processing that were able to do and radically reduce the hardware footprint of the systems that were previously in place to be able to do the processing that we needed to do specifically related to quality reporting the financial analysis decision support pricing risk analysis all of those kind of functions and i will say in general we’ve been very pleased with what we’ve been able to accomplish in a short period of time a little over a year of actual implementation has gotten us like I said about seventy three percent faster on our ability to produce reports to do the analytics required to respond to our state and federal partners and as well as to effectively closed our books meet our monthly financial reporting obligations those kinds of things and that I think concludes my selection of slides thanks so much James so just a reminder please feel free to send your questions in during during this Q&A part using the Q&A console at the bottom of the screen we’ll do our best to answer them so James I think we’ve got our first question for you can you expand on how well care was able to accelerate reporting schedules from 30 days down to eight days sure so the the previous technology we had in place were kind of a hodgepodge of multiple systems that included manipulating data from various transactional systems that were running on Oracle Microsoft sequel server and my sequel the first step in that process was to homogenize that data get everything loaded into our oracle enterprise data warehouse and then begin the process of doing multiple stages of analytics data cleansing codification those kinds of things to pull in a result set that would allow us to do our our deep analytics on it the first step in that process for us to be able to speed that up was to use attunity replicate to do change data capture and be able to load all of the various data sources that we have from those technology as I mentioned or cold sequel server my sequel and bring those straight into pivotal Greenplum that gave us the just a huge step in using things like our man custom ETL custom PL sequel to do a lot of manipulation on it before we were ever able to do analytics once we get the data into into pivotal Greenplum were able to quickly establish relationships and begin our analysis very close to real time like we as we were talking about in the presentation one of the big things for us was you know once we saw the data set it basically took us a month to see the data set before we were able to do the analytics on it now we’re able to really look at on a daily basis fresh data look at the way things are trending be able to you know more quickly produce that result great so our next question I think is for Dave how is the partnership with attunity helping other pivotal customers I think many of our pivotal customers go through the same exercise the change is describing and Kevin described in terms of the technology so generally when we first are engaged with the tuner off and I would say the customer is performing an initial load of the pivotal technologies right so they’re they’re using other technologies

like James mentioned and they want to evaluate or have made a decision to use the pivotal technologies and so the first step you have to do is get the data into the pivotal technology and so you know they’ll use the transfer mechanisms to create those initial populations of the databases and then once the technology is in place they’re moving forward on a regular basis i talked about operationalizing activities so one of the steps in operationalizing is to tin you to load the changes to the source databases capture those changes the change data capture mechanism and take that change data and add it to the data that has already been loaded into other pivotal technology so generally we see customers you know moving through that progression and and we also see scenarios where once they’ve had a success in one part of the business then perhaps they want to expand into other parts of the business and so we’ll see the relationship with attunity you know following us around the organization as we go to new groups within the customer organizations they’ll also adopt the attunity technology to go through the same process in those groups great and for Kevin we have a question would you give us some other examples of health care companies they’re using attunity software to optimize and integrate data chart great question we have a number of plants that use us in order to feed data into or out of the epic system so we’ve got one health care provider that uses that book they feed data changes to the underlying article database what they’re specifically using our replicate software to do is to run it from clarity which is a part of the epic system to teradata and to do that to support reporting so that’s just one of several examples in which the the epic system we’re fully compliant with the the underlying database moving data into and out of it super and so I think this question is is mainly for Davis that maybe also Kevin can chime in this person says if we have hortonworks dupe and green plum do we really need Hawk in place well the answer is maybe so like I described we provide a flexible license that allows you to utilize different components you know if and when you need them greenplum certainly has the ability to access data that’s in Hadoop it does it a little less directly than honk so if you wanted to utilize hawk the reason you would probably choose to utilize hawk is if the volume of data that you want to access is so large that you don’t want to spend the time to move it into greenville um that that would be a reason or if the performance want to you know access the data directly and you want to have the queries operating directly on top of the original source of the data that might be a reason to do so it’s not a requirement we do support Hortonworks as a underlying store for the hawk capabilities its announcement we made earlier this year we form something called the open data platform together with IBM important works others are certainly welcome to join but those are the folks who have joined this OOP distribution vendors who’ve joined so far and what that means is that we can interoperate with those other Hadoop platforms so the customer is running the hawk capabilities on top before it works but but it’s entirely up to you if you know if you’re satisfied with the architecture the way it works today then I would say no you don’t need to add the heart but you do have the option great so this is a question for eternity person asks what rfid tools or technology are you using to track the movement of equipment in hospitals sure that i was speaking on behalf of a client and they would be more equipped to provide details on that but i think the the key point would be that by aggregating data about the movement and the transactions if you will of various pieces of equipment individuals they were able to get a great operational flow of their system over a 30 day period we can follow up as to which specific implementation of RFID was used in that case okay this question just came in this F how do you compare the solution with Impala sure I’ll take that Kathryn Dave mom so I specifically

highlighted T pcds because I thought that’s a good representation of how it differs from Impala so first of all Impala is Claude eras sequel capability built on top of a doob so the Hadoop industry has recognized in general that having sequel capabilities on top of the dupe is a good thing all the major who do providers are you know adding sequel capabilities we’ve chosen to do it by taking an existing sequel implementation in plum database and offering it on top of the dude so what that meant was we already had a very rich set of sequel capabilities most of the other approaches are starting from the bottom up so building sequel processing parallel sequel processing from the ground up on top of their Hadoop Impala implementation representin we recently performed some analysis and learned that among the different the hawk implication is the most complete and so that T BCBS benchmark as I said has 111 query form all 100 Impala as of our last testing could only do 30 one of those queries now I’m sure they’ll be you know knocking more and more of those off but it’s a long you know slow process to get that complete of Syria query session capable so the differentiate we both agree that sequel I medupe is important the differentiation right now is that for some period of time you know we’ll have a significant lead sequel in the performance marsh equal to do great and this question comes in for Kevin is your CDC technology managed by triggers or is it log based a great question we are log based we don’t require any software to be installed on the source of the target so we have a very low footprint and we are based on logs and I think this might be a follow-up as well as how do you schedule eternity does it run all the time is that the same question I wasn’t sure it does run on a continuous basis okay I mean it certainly and if you’re in using the change data capture technology which can send continuous somewhat continuous data streams based on whatever increments work best for the business right and then one more for cabin does the Trinity also supports sequel server and db2 s sources we do we support sequel server and db2 is both sources and targets so as I mentioned before we support actually I think the numbers over 35 now sources and targets and so that includes all the major data warehouse all the major database platforms on the cloud side we support AWS and azure which we announced at this month we support MongoDB in the no sequel camp and then for a dupe all the major Hadoop distributions most notably hortonworks it by extension of it’ll HD the we exist is a Switzerland between all those major platforms okay and then I think we had a follow-up to the previous question which is can attunity run on a scheduler and I know if this is already answered if this is something new I don’t have a specific answer that question we can follow up on that okay sounds good well I think we are just about out of questions so with that will bring this webinar to a close we’ll be sending out the links to the recording and a SlideShare deck of the session within the next few days there’s also more information on the pivotal attunity and well-cared websites if you’re interested thank you again for joining us and have a great day