>> Paige: Hello everybody and thank you for joining us today for the virtual Vertica BDC 2020. Today's breakout session is entitled Vertica in Eon Mode past, present and future. I'm Paige Roberts, open source relations manager at Vertica and I'll be your host for this session. Joining me is Vertica engineer, Yuanzhe Bei and Vertica Product Manager, David Sprogis. Before we begin, I encourage you to submit questions or comments during the virtual session. You don't have to wait till the end. Just type your question or comment as you think of it in the question box, below the slides and click Submit. Q&A session at the end of the presentation. We'll answer as many of your questions as we're able to during that time, and any questions that we don't address, we'll do our best to answer offline. If you wish after the presentation, you can visit the Vertica forums to post your questions there and our engineering team is planning to join the forums to keep the conversation going, just like a Dev Lounge at a normal in person, BDC. So, as a reminder, you can maximize your screen by clicking the double arrow button in the lower right corner of the slides, if you want to see them bigger. And yes, before you ask, this virtual session is being recorded and will be available to view on demand this week. We are supposed to send you a notification as soon as it's ready. All right, let's get started. Over to you, Dave. >> David: Thanks, Paige. Hey, everybody. Let's start with a timeline of the life of Eon Mode. About two years ago, a little bit less than two years ago, we introduced Eon Mode on AWS. Pretty specifically for the purpose of rapid scaling to meet the cloud economics promise. It wasn't long after that we realized that workload isolation, a byproduct of the architecture was very important to our users and going to the third tick, you can see that the importance of that workload isolation was manifest in Eon Mode being made available on-premise using Pure Storage FlashBlade. Moving to the fourth tick mark, we took steps to improve workload isolation, with a new type of subcluster which Yuanzhe will go through and to the fifth tick mark, the introduction of secondary subclusters for faster scaling and other improvements which we will cover in the slides to come. Getting started with, why we created Eon Mode in the first place. Let's imagine that your database is this pie, the pecan pie and we're loading pecan data in through the ETL cutting board in the upper left hand corner. We have a couple of free floating pecans, which we might imagine to be data supporting external tables. As you know, the Vertica has a query engine capability as well which we call external tables. And so if we imagine this pie, we want to serve it with a number of servers. Well, let's say we wanted to serve it with three servers, three nodes, we would need to slice that pie into three segments and we would serve each one of those segments from one of our nodes. Now because the data is important to us and we don't want to lose it, we're going to be saving that data on some kind of raid storage or redundant storage. In case one of the drives goes bad, the data remains available because of the durability of raid. Imagine also, that we care about the availability of the overall database. Imagine that a node goes down, perhaps the second node goes down, we still want to be able to query our data and through nodes one and three, we still have all three shards covered and we can do this because of buddy projections. Each neighbor, each nodes neighbor contains a copy of the data from the node next to it. And so in this case, node one is sharing its segment with node two. So node two can cover node one, node three can cover node two and node one back to node three. Adding a little bit more complexity, we might store the data in different copies, each copy sorted for a different kind of query. We call this projections in Vertica and for each projection, we have another copy of the data sorted differently. Now it gets complex. What happens when we want to add a node? Well, if we wanted to add a fourth node here, what we would have to do, is figure out how to re-slice all of the data in all of the copies that we have. In effect, what we want to do is take our three slices and slice it into four, which means taking a portion of each of our existing thirds and re-segmenting into quarters. Now that looks simple in the graphic here, but when it comes to moving data around, it becomes quite complex because for each copy of each segment we need to replace it and move that data on to the new node. What's more, the fourth node can't have a copy of itself that would be problematic in case it went down. Instead, what we need is we need that buddy to be sitting on another node, a neighboring node. So we need to re-orient the buddies as well. All of this takes a lot of time, it can take 12, 24 or even 36 hours in a period when you do not want your database under high demand. In fact, you may want to stop loading data altogether in order to speed it up. This is a planned event and your applications should probably be down during this period, which makes it difficult. With the advent of cloud computing, we saw that services were coming up and down faster and we determined to re-architect Vertica in a way to accommodate that rapid scaling. Let's see how we did it. So let's start with four nodes now and we've got our four nodes database. Let's add communal storage and move each of the segments of data into communal storage. Now that's the separation that we're talking about. What happens if we run queries against it? Well, it turns out that the communal storage is not necessarily performing and so the IO would be slow, which would make the overall queries slow. In order to compensate for the low performance of communal storage, we need to add back local storage, now it doesn't have to be raid because this is just an ephemeral copy but with the data files, local to the node, the queries will run much faster. In AWS, communal storage really does mean an S3 bucket and here's a simplified version of the diagram. Now, do we need to store all of the data from the segment in the depot? The answer is no and the graphics inside the bucket has changed to reflect that. It looks more like a bullseye, showing just a segment of the data being copied to the cache or to the depot, as we call it on each one of the nodes. How much data do you store on the node? Well, it would be the active data set, the last 30 days, the last 30 minutes or the last. Whatever period of time you're working with. The active working set is the hot data and that's how large you want to size your depot. By architecting this way, when you scale up, you're not re-segmenting the database. What you're doing, is you're adding more compute and more subscriptions to the existing shards of the existing database. So in this case, we've added a complete set of four nodes. So we've doubled our capacity and we've doubled our subscriptions, which means that now, the two nodes can serve the yellow shard, two nodes can serve the red shard and so on. In this way, we're able to run twice as many queries in the same amount of time. So you're doubling the concurrency. How high can you scale? Well, can you scale to 3X, 5X? We tested this in the graphics on the right, which shows concurrent users in the X axis by the number of queries executed in a minute along the Y axis. We've grouped execution in runs of 10 users, 30 users, 50, 70 up to 150 users. Now focusing on any one of these groups, particularly up around 150. You can see through the three bars, starting with the bright purple bar, three nodes and three segments. That as you add nodes to the middle purple bar, six nodes and three segments, you've almost doubled your throughput up to the dark purple bar which is nine nodes and three segments and our tests show that you can go to 5X with pretty linear performance increase. Beyond that, you do continue to get an increase in performance but your incremental performance begins to fall off. Eon architecture does something else for us and that is it provides high availability because each of the nodes can be thought of as ephemeral and in fact, each node has a buddy subscription in a way similar to the prior architecture. So if we lose node four, we're losing the node responsible for the red shard and now node one has to pick up responsibility for the red shard while that node is down. When a query comes in, and let's say it comes into one and one is the initiator then one will look for participants, it'll find a blue shard and a green shard but when it's looking for the red, it finds itself and so the node number one will be doing double duty. This means that your performance will be cut in half approximately, for the query. This is acceptable until you are able to restore the node. Once you restore it and once the depot becomes rehydrated, then your performance goes back to normal. So this is a much simpler way to recover nodes in the event of node failure. By comparison, Enterprise Mode the older architecture. When we lose the fourth node, node one takes over responsibility for the first shard and the yellow shard and the red shard. But it also is responsible for rehydrating the entire data segment of the red shard to node four, this can be very time consuming and imposes even more stress on the first node. So performance will go down even further. Eon Mode has another feature and that is you can scale down completely to zero. We call this hibernation, you shut down your database and your database will maintain full consistency in a rest state in your S3 bucket and then when you need access to your database again, you simply recreate your cluster and revive your database and you can access your database once again. That concludes the rapid scaling portion of, why we created Eon Mode. To take us through workload isolation is Yuanzhe Bei, Yuanzhe. >> Yuanzhe: Thanks Dave, for presenting how Eon works in general. In the next section, I will show you another important capability of Vertica Eon Mode, the workload isolation. Dave used a pecan pie as an example of database. Now let's say it's time for the main course. Does anyone still have a problem with food touching on their plates. Parents know that it's a common problem for kids. Well, we have a similar problem in database as well. So there could be multiple different workloads accessing your database at the same time. Say you have ETL jobs running regularly. While at the same time, there are dashboards running short queries against your data. You may also have the end of month report running and their can be ad hoc data scientists, connect to the database and do whatever the data analysis they want to do and so on. How to make these mixed workload requests not interfere with each other is a real challenge for many DBAs. Vertica Eon Mode provides you the solution. I'm very excited here to introduce to you to the important concept in Eon Mode called subclusters. In Eon Mode, nodes they belong to the predefined subclusters rather than the whole cluster. DBAs can define different subcluster for different kinds of workloads and it redirects those workloads to the specific subclusters. For example, you can have an ETL subcluster, dashboard subcluster, report subcluster and the analytic machine learning subcluster. Vertica Eon subcluster is designed to achieve the three main goals. First of all, strong workload isolation. That means any operation in one subcluster should not affect or be affected by other subclusters. For example, say the subcluster running the report is quite overloaded and already there can be, the data scienctists running crazy analytic jobs, machine learning jobs on the analytics subcluster and making it very slow, even stuck or crash or whatever. In such scenario, your ETL and dashboards subcluster should not be or at least very minimum be impacted by this crisis and which means your ETL job which should not lag behind and dashboard should respond timely. We have done a lot of improvements as of 10.0 release and will continue to deliver improvements in this category. Secondly, fully customized subcluster settings. That means any subcluster can be set up and tuned for very different workloads without affecting other subclusters. Users should be able to tune up, tune down, certain parameters based on the actual needs of the individual subcluster workload requirements. As of today, Vertica already supports few settings that can be done at the subcluster level for example, the depot pinning policy and then we will continue extending more that is like resource pools (mumbles) in the near future. Lastly, Vertica subclusters should be easy to operate and cost efficient. What it means is that the subcluster should be able to turn on, turn off, add or remove or should be available for use according to rapid changing workloads. Let's say in this case, you want to spin up more dashboard subclusters because we need higher scores report, we can do that. You might need to run several report subclusters because you might want to run multiple reports at the same time. While on the other hand, you can shut down your analytic machine learning subcluster because no data scientists need to use it at this moment. So we made automate a lot of change, the improvements in this category, which I'll explain in detail later and one of the ultimate goal is to support auto scaling To sum up, what we really want to deliver for subcluster is very simple. You just need to remember that accessing subclusters should be just like accessing individual clusters. Well, these subclusters do share the same catalog. So you don't have to work out the stale data and don't need to worry about data synchronization. That'd be a nice goal, Vertica upcoming 10.0 release is certainly a milestone towards that goal, which will deliver a large part of the capability in this direction and then we will continue to improve it after 10.0 release. In the next couple of slides, I will highlight some issues about workload isolation in the initial Eon release and show you how we resolve these issues. First issue when we initially released our first or so called subcluster mode, it was implemented using fault groups. Well, fault groups and the subcluster have something in common. Yes, they are both defined as a set of nodes. However, they are very different in all the other ways. So, that was very confusing in the first place, when we implement this. As of 9.3.0 version, we decided to detach subcluster definition from the fault groups, which enabled us to further extend the capability of subclusters. Fault groups in the pre 9.3.0 versions will be converted into subclusters during the upgrade and this was a very important step that enabled us to provide all the amazing, following improvements on subclusters. The second issue in the past was that it's hard to control the execution groups for different types of workloads. There are two types of problems here and I will use some example to explain. The first issue is about control group size. There you allocate six nodes for your dashboard subcluster and what you really want is on the left, the three pairs of nodes as three execution groups, and each pair of nodes will need to subscribe to all the four shards. However, that's not really what you get. What you really get is there on the right side that the first four nodes subscribed to one shard each and the rest two nodes subscribed to two dangling shards. So you won't really get three execusion groups but instead only get one and two extra nodes have no value at all. The solution is to use subclusters. So instead of having a subcluster with six nodes, you can split it up into three smaller ones. Each subcluster will guarantee to subscribe to all the shards and you can further handle this three subcluster using load balancer across them. In this way you achieve the three real exclusion groups. The second issue is that the session participation is non-deterministic. Any session will just pick four random nodes from the subcluster as long as this covers one shard each. In other words, you don't really know which set of nodes will make up your execution group. What's the problem? So in this case, the fourth node will be doubled booked by two concurrent sessions. And you can imagine that the resource usage will be imbalanced and both queries performance will suffer. What is even worse is that these queries of the two concurrent sessions target different table They will cause the issue, that depot efficiency will be reduced, because both session will try to fetch the files on to two tables into the same depot and if your depot is not large enough, they will evict each other, which will be very bad. To solve this the same way, you can solve this by declaring subclusters, in this case, two subclusters and a load balancer group across them. The reason it solved the problem is because the session participation would not go across the boundary. So there won't be a case that any node is double booked and in terms of the depot and if you use the subcluster and avoid using a load balancer group, and carefully send the first workload to the first subcluster and the second to the second subcluster and then the result is that depot isolation is achieved. The first subcluster will maintain the data files for the first query and you don't need to worry about the file being evicted by the second kind of session. Here comes the next issue, it's the scaling down. In the old way of defining subclusters, you may have several execution groups in the subcluster. You want to shut it down, one or two execution groups to save cost. Well, here comes the pain, because you don't know which nodes may be used by which session at any point, it is hard to find the right timing to hit the shutdown button of any of the instances. And if you do and get unlucky, say in this case, you pull the first four nodes, one of the session will fail because it's participating in the node two and node four at that point. User of that session will notice because their query fails and we know that for many business this is critical problem and not acceptable. Again, with subclusters this problem is resolved. Same reason, session cannot go across the subcluster boundary. So all you need to do is just first prevent query sent to the first subcluster and then you can shut down the instances in that subcluster. You are guaranteed to not break any running sessions. Now, you're happy and you want to shut down more subclusters then you hit the issue four, the whole cluster will go down, why? Because the cluster loses quorum. As a distributed system, you need to have at least more than half of a node to be up in order to commit and keep the cluster up. This is to prevent the catalog diversion from happening, which is important. But do you still want to shut down those nodes? Because what's the point of keeping those nodes up and if you are not using them and let them cost you money right. So Vertica has a solution, you can define a subcluster as secondary to allow them to shut down without worrying about quorum. In this case, you can define the first three subclusters as secondary and the fourth one as primary. By doing so, this secondary subclusters will not be counted towards the quorum because we changed the rule. Now instead of requiring more than half of node to be up, it only require more than half of the primary node to be up. Now you can shut down your second subcluster and even shut down your third subcluster as well and keep the remaining primary subcluster to be still running healthily. There are actually more benefits by defining secondary subcluster in addition to the quorum concern, because the secondary subclusters no longer have the voting power, they don't need to persist catalog anymore. This means those nodes are faster to deploy, and can be dropped and re-added. Without the worry about the catalog persistency. For the most the subcluster that only need to read only query, it's the best practice to define them as secondary. The commit will be faster on this secondary subcluster as well, so running this query on the secondary subcluster will have less spikes. Primary subcluster as usual handle everything is responsible for consistency, the background tasks will be running. So DBAs should make sure that the primary subcluster is stable and assume is running all the time. Of course, you need to at least one primary subcluster in your database. Now with the secondary subcluster, user can start and stop as they need, which is very convenient and this further brings up another issue is that if there's an ETL transaction running and in the middle, a subcluster starting and it become up. In older versions, there is no catalog resync mechanism to keep the new subcluster up to date. So Vertica rolls back to ETL session to keep the data consistency. This is actually quite disruptive because real world ETL workloads can sometimes take hours and rolling back at the end means, a large waste of resources. We resolved this issue in 9.3.1 version by introducing a catalog resync mechanism when such situation happens. ETL transactions will not roll back anymore, but instead will take some time to resync the catalog and commit and the problem is resolved. And last issue I would like to talk about is the subscription. Especially for large subcluster when you start it, the startup time is quite long, because the subscription commit used to be serialized. In one of the in our internal testing with large catalogs committing a subscription, you can imagine it takes five minutes. Secondary subcluster is better, because it doesn't need to persist the catalog during the commit but still take about two seconds to commit. So what's the problem here? Let's do the math and look at this chart. The X axis is the time in the minutes and the Y axis is the number of nodes to be subscribed. The dark blues represents your primary subcluster and light blue represents the secondary subcluster. Let's say the subcluster have 16 nodes in total and if you start a secondary subcluster, it will spend about 30 seconds in total, because the 2 seconds times 16 is 32. It's not actually that long time. but if you imagine that starting secondary subcluster, you expect it to be super fast to react to the fast changing workload and 30 seconds is no longer trivial anymore and what is even worse is on the primary subcluster side. Because the commit is much longer than five minutes let's assume, then at the point, you are committing to six nodes subscription all other nodes already waited for 30 minutes for GCLX or we know the global catalog lock, and the Vertica will crash the nodes, if any node cannot get the GCLX for 30 minutes. So the end result is that your whole database crashed. That's a serious problem and we know that and that's why we are already planning for the fix, for the 10.0, so that all the subscription will be batched up and all the nodes will commit at the same time concurrently. And by doing that, you can imagine the primary subcluster can finish commiting in five minutes instead of crashing and the secondary subcluster can be finished even in seconds. That summarizes the highlights for the improvements we have done as of 10.0, and I hope you already get excited about Emerging Eon Deployment Pattern that's shown here. A primary subcluster that handles data loading, ETL jobs and tuple mover jobs is the backbone of the database and you keep it running all the time. At the same time defining different secondary subcluster for different workloads and provision them when the workload requirement arrives and then de-provision them when the workload is done to save the operational cost. So can't wait to play with the subcluster. Here as are some Admin Tools command you can start using. And for more details, check out our Eon subcluster documentation for more details. And thanks everyone for listening and I'll head back to Dave to talk about the Eon on-prem. >> David: Thanks Yuanzhe. At the same time that Yuanzhe and the rest of the dev team were working on the improvements that Yuanzhe described in and other improvements. This guy, John Yovanovich, stood on stage and told us about his deployment at at&t where he was running Eon Mode on-prem. Now this was only six months after we had launched Eon Mode on AWS. So when he told us that he was putting it into production on-prem, we nearly fell out of our chairs. How is this possible? We took a look back at Eon and determined that the workload isolation and the improvement to the operations for restoring nodes and other things had sufficient value that John wanted to run it on-prem. And he was running it on the Pure Storage FlashBlade. Taking a second look at the FlashBlade we thought alright well, does it have the performance? Yes, it does. The FlashBlade is a collection of individual blades, each one of them with NVMe storage on it, which is not only performance but it's scalable and so, we then asked is it durable? The answer is yes. The data safety is implemented with the N+2 redundancy which means that up to two blades can fail and the data remains available. And so with this we realized DBAs can sleep well at night, knowing that their data is safe, after all Eon Mode outsources the durability to the communal storage data store. Does FlashBlade have the capacity for growth? Well, yes it does. You can start as low as 120 terabytes and grow as high as about eight petabytes. So it certainly covers the range for most enterprise usages. And operationally, it couldn't be easier to use. When you want to grow your database. You can simply pop new blades into the FlashBlade unit, and you can do that hot. If one goes bad, you can pull it out and replace it hot. So you don't have to take your data store down and therefore you don't have to take Vertica down. Knowing all of these things we got behind Pure Storage and partnered with them to implement the first version of Eon on-premise. That changed our roadmap a little bit. We were imagining it would start with Amazon and then go to Google and then to Azure and at some point to Alibaba cloud, but as you can see from the left column, we started with Amazon and went to Pure Storage. And then from Pure Storage, we went to Minio and we launched Eon Mode on Minio at the end of last year. Minio is a little bit different than Pure Storage. It's software only, so you can run it on pretty much any x86 servers and you can cluster them with storage to serve up an S3 bucket. It's a great solution for up to about 120 terabytes Beyond that, we're not sure about performance implications cause we haven't tested it but for your dev environments or small production environments, we think it's great. With Vertica 10, we're introducing Eon Mode on Google Cloud. This means not only running Eon Mode in the cloud, but also being able to launch it from the marketplace. We're also offering Eon Mode on HDFS with version 10. If you have a Hadoop environment, and you want to breathe new fresh life into it with the high performance of Vertica, you can do that starting with version 10. Looking forward we'll be moving Eon mode to Microsoft Azure. We expect to have something breathing in the fall and offering it to select customers for beta testing and then we expect to release it sometime in 2021 Following that, further on horizon is Alibaba cloud. Now, to be clear we will be putting, Vertica in Enterprise Mode on Alibaba cloud in 2020 but Eon Mode is going to trail behind whether it lands in 2021 or not, we're not quite sure at this point. Our goal is to deliver Eon Mode anywhere you want to run it, on-prem or in the cloud, or both because that is one of the great value propositions of Vertica is the hybrid capability, the ability to run in both your on prem environment and in the cloud. What's next, I've got three priority and roadmap slides. This is the first of the three. We're going to start with improvements to the core of Vertica. Starting with query crunching, which allows you to run long running queries faster by getting nodes to collaborate, you'll see that coming very soon. We'll be making improvements to large clusters and specifically large cluster mode. The management of large clusters over 60 nodes can be tedious. We intend to improve that. In part, by creating a third network channel to offload some of the communication that we're now loading onto our spread or agreement protocol. We'll be improving depot efficiency. We'll be pushing down more controls to the subcluster level, allowing you to control your resource pools at the subcluster level and we'll be pairing tuple moving with data loading. From an operational flexibility perspective, we want to make it very easy to shut down and revive primaries and secondaries on-prem and in the cloud. Right now, it's a little bit tedious, very doable. We want to make it as easy as a walk in the park. We also want to allow you to be able to revive into a different size subcluster and last but not least, in fact, probably the most important, the ability to change shard count. This has been a sticking point for a lot of people and it puts a lot of pressure on the early decision of how many shards should my database be? Whether it's in 2020 or 2021. We know it's important to you so it's important to us. Ease of use is also important to us and we're making big investments in the management console, to improve managing subclusters, as well as to help you manage your load balancer groups. We also intend to grow and extend Eon Mode to new environments. Now we'll take questions and answers

>> Joy: Welcome to the Virtual Big Data Conference. Vertica is so excited to host this event. I'm Joy King, and I'll be your host for today's Big Data Conference Keynote Session. It's my honor and my genuine pleasure to lead Vertica's product and go-to-market strategy. And I'm so lucky to have a passionate and committed team who turned our Vertica BDC event, into a virtual event in a very short amount of time. I want to thank the thousands of people, and yes, that's our true number who have registered to attend this virtual event. We were determined to balance your health, safety and your peace of mind with the excitement of the Vertica BDC. This is a very unique event. Because as I hope you all know, we focus on engineering and architecture, best practice sharing and customer stories that will educate and inspire everyone. I also want to thank our top sponsors for the virtual BDC, Arrow, and Pure Storage. Our partnerships are so important to us and to everyone in the audience. Because together, we get things done faster and better. Now for today's keynote, you'll hear from three very important and energizing speakers. First, Colin Mahony, our SVP and General Manager for Vertica, will talk about the market trends that Vertica is betting on to win for our customers. And he'll share the exciting news about our Vertica 10 announcement and how this will benefit our customers. Then you'll hear from Amy Fowler, VP of strategy and solutions for FlashBlade at Pure Storage. Our partnership with Pure Storage is truly unique in the industry, because together modern infrastructure from Pure powers modern analytics from Vertica. And then you'll hear from John Yovanovich, Director of IT at AT&T, who will tell you about the Pure Vertica Symphony that plays live every day at AT&T. Here we go, Colin, over to you. >> Colin: Well, thanks a lot joy. And, I want to echo Joy's thanks to our sponsors, and so many of you who have helped make this happen. This is not an easy time for anyone. We were certainly looking forward to getting together in person in Boston during the Vertica Big Data Conference and Winning with Data. But I think all of you and our team have done a great job, scrambling and putting together a terrific virtual event. So really appreciate your time. I also want to remind people that we will make both the slides and the full recording available after this. So for any of those who weren't able to join live, that is still going to be available. Well, things have been pretty exciting here. And in the analytic space in general, certainly for Vertica, there's a lot happening. There are a lot of problems to solve, a lot of opportunities to make things better, and a lot of data that can really make every business stronger, more efficient, and frankly, more differentiated. For Vertica, though, we know that focusing on the challenges that we can directly address with our platform, and our people, and where we can actually make the biggest difference is where we ought to be putting our energy and our resources. I think one of the things that has made Vertica so strong over the years is our ability to focus on those areas where we can make a great difference. So for us as we look at the market, and we look at where we play, there are really three recent and some not so recent, but certainly picking up a lot of the market trends that have become critical for every industry that wants to Win Big With Data. We've heard this loud and clear from our customers and from the analysts that cover the market. If I were to summarize these three areas, this really is the core focus for us right now. We know that there's massive data growth. And if we can unify the data silos so that people can really take advantage of that data, we can make a huge difference. We know that public clouds offer tremendous advantages, but we also know that balance and flexibility is critical. And we all need the benefit that machine learning for all the types up to the end data science. We all need the benefits that they can bring to every single use case, but only if it can really be operationalized at scale, accurate and in real time. And the power of Vertica is, of course, how we're able to bring so many of these things together. Let me talk a little bit more about some of these trends. So one of the first industry trends that we've all been following probably now for over the last decade, is Hadoop and specifically HDFS. So many companies have invested, time, money, more importantly, people in leveraging the opportunity that HDFS brought to the market. HDFS is really part of a much broader storage disruption that we'll talk a little bit more about, more broadly than HDFS. But HDFS itself was really designed for petabytes of data, leveraging low cost commodity hardware and the ability to capture a wide variety of data formats, from a wide variety of data sources and applications. And I think what people really wanted, was to store that data before having to define exactly what structures they should go into. So over the last decade or so, the focus for most organizations is figuring out how to capture, store and frankly manage that data. And as a platform to do that, I think, Hadoop was pretty good. It certainly changed the way that a lot of enterprises think about their data and where it's locked up. In parallel with Hadoop, particularly over the last five years, Cloud Object Storage has also given every organization another option for collecting, storing and managing even more data. That has led to a huge growth in data storage, obviously, up on public clouds like Amazon and their S3, Google Cloud Storage and Azure Blob Storage just to name a few. And then when you consider regional and local object storage offered by cloud vendors all over the world, the explosion of that data, in leveraging this type of object storage is very real. And I think, as I mentioned, it's just part of this broader storage disruption that's been going on. But with all this growth in the data, in all these new places to put this data, every organization we talk to is facing even more challenges now around the data silo. Sure the data silos certainly getting bigger. And hopefully they're getting cheaper per bit. But as I said, the focus has really been on collecting, storing and managing the data. But between the new data lakes and many different cloud object storage combined with all sorts of data types from the complexity of managing all this, getting that business value has been very limited. This actually takes me to big bet number one for Team Vertica, which is to unify the data. Our goal, and some of the announcements we have made today plus roadmap announcements I'll share with you throughout this presentation. Our goal is to ensure that all the time, money and effort that has gone into storing that data, all the data turns into business value. So how are we going to do that? With a unified analytics platform that analyzes the data wherever it is HDFS, Cloud Object Storage, External tables in an any format ORC, Parquet, JSON, and of course, our own Native Roth Vertica format. Analyze the data in the right place in the right format, using a single unified tool. This is something that Vertica has always been committed to, and you'll see in some of our announcements today, we're just doubling down on that commitment. Let's talk a little bit more about the public cloud. This is certainly the second trend. It's the second wave maybe of data disruption with object storage. And there's a lot of advantages when it comes to public cloud. There's no question that the public clouds give rapid access to compute storage with the added benefit of eliminating data center maintenance that so many companies, want to get out of themselves. But maybe the biggest advantage that I see is the architectural innovation. The public clouds have introduced so many methodologies around how to provision quickly, separating compute and storage and really dialing-in the exact needs on demand, as you change workloads. When public clouds began, it made a lot of sense for the cloud providers and their customers to charge and pay for compute and storage in the ratio that each use case demanded. And I think you're seeing that trend, proliferate all over the place, not just up in public cloud. That architecture itself is really becoming the next generation architecture for on-premise data centers, as well. But there are a lot of concerns. I think we're all aware of them. They're out there many times for different workloads, there are higher costs. Especially if some of the workloads that are being run through analytics, which tend to run all the time. Just like some of the silo challenges that companies are facing with HDFS, data lakes and cloud storage, the public clouds have similar types of siloed challenges as well. Initially, there was a belief that they were cheaper than data centers, and when you added in all the costs, it looked that way. And again, for certain elastic workloads, that is the case. I don't think that's true across the board overall. Even to the point where a lot of the cloud vendors aren't just charging lower costs anymore. We hear from a lot of customers that they don't really want to tether themselves to any one cloud because of some of those uncertainties. Of course, security and privacy are a concern. We hear a lot of concerns with regards to cloud and even some SaaS vendors around shared data catalogs, across all the customers and not enough separation. But security concerns are out there, you can read about them. I'm not going to jump into that bandwagon. But we hear about them. And then, of course, I think one of the things we hear the most from our customers, is that each cloud stack is starting to feel even a lot more locked in than the traditional data warehouse appliance. And as everybody knows, the industry has been running away from appliances as fast as it can. And so they're not eager to get locked into another, quote, unquote, virtual appliance, if you will, up in the cloud. They really want to make sure they have flexibility in which clouds, they're going to today, tomorrow and in the future. And frankly, we hear from a lot of our customers that they're very interested in eventually mixing and matching, compute from one cloud with, say storage from another cloud, which I think is something that we'll hear a lot more about. And so for us, that's why we've got our big bet number two. we love the cloud. We love the public cloud. We love the private clouds on-premise, and other hosting providers. But our passion and commitment is for Vertica to be able to run in any of the clouds that our customers choose, and make it portable across those clouds. We have supported on-premises and all public clouds for years. And today, we have announced even more support for Vertica in Eon Mode, the deployment option that leverages the separation of compute from storage, with even more deployment choices, which I'm going to also touch more on as we go. So super excited about our big bet number two. And finally as I mentioned, for all the hype that there is around machine learning, I actually think that most importantly, this third trend that team Vertica is determined to address is the need to bring business critical, analytics, machine learning, data science projects into production. For so many years, there just wasn't enough data available to justify the investment in machine learning. Also, processing power was expensive, and storage was prohibitively expensive. But to train and score and evaluate all the different models to unlock the full power of predictive analytics was tough. Today you have those massive data volumes. You have the relatively cheap processing power and storage to make that dream a reality. And if you think about this, I mean with all the data that's available to every company, the real need is to operationalize the speed and the scale of machine learning so that these organizations can actually take advantage of it where they need to. I mean, we've seen this for years with Vertica, going back to some of the most advanced gaming companies in the early days, they were incorporating this with live data directly into their gaming experiences. Well, every organization wants to do that now. And the accuracy for clickability and real time actions are all key to separating the leaders from the rest of the pack in every industry when it comes to machine learning. But if you look at a lot of these projects, the reality is that there's a ton of buzz, there's a ton of hype spanning every acronym that you can imagine. But most companies are struggling, do the separate teams, different tools, silos and the limitation that many platforms are facing, driving, down sampling to get a small subset of the data, to try to create a model that then doesn't apply, or compromising accuracy and making it virtually impossible to replicate models, and understand decisions. And if there's one thing that we've learned when it comes to data, prescriptive data at the atomic level, being able to show end of one as we refer to it, meaning individually tailored data. No matter what it is healthcare, entertainment experiences, like gaming or other, being able to get at the granular data and make these decisions, make that scoring applies to machine learning just as much as it applies to giving somebody a next-best-offer. But the opportunity has never been greater. The need to integrate this end-to-end workflow and support the right tools without compromising on that accuracy. Think about it as no downsampling, using all the data, it really is key to machine learning success. Which should be no surprise then why the third big bet from Vertica is one that we've actually been working on for years. And we're so proud to be where we are today, helping the data disruptors across the world operationalize machine learning. This big bet has the potential to truly unlock, really the potential of machine learning. And today, we're announcing some very important new capabilities specifically focused on unifying the work being done by the data science community, with their preferred tools and platforms, and the volume of data and performance at scale, available in Vertica. Our strategy has been very consistent over the last several years. As I said in the beginning, we haven't deviated from our strategy. Of course, there's always things that we add. Most of the time, it's customer driven, it's based on what our customers are asking us to do. But I think we've also done a great job, not trying to be all things to all people. Especially as these hype cycles flare up around us, we absolutely love participating in these different areas without getting completely distracted. I mean, there's a variety of query tools and data warehouses and analytics platforms in the market. We all know that. There are tools and platforms that are offered by the public cloud vendors, by other vendors that support one or two specific clouds. There are appliance vendors, who I was referring to earlier who can deliver package data warehouse offerings for private data centers. And there's a ton of popular machine learning tools, languages and other kits. But Vertica is the only advanced analytic platform that can do all this, that can bring it together. We can analyze the data wherever it is, in HDFS, S3 Object Storage, or Vertica itself. Natively we support multiple clouds on-premise deployments, And maybe most importantly, we offer that choice of deployment modes to allow our customers to choose the architecture that works for them right now. It still also gives them the option to change move, evolve over time. And Vertica is the only analytics database with end-to-end machine learning that can truly operationalize ML at scale. And I know it's a mouthful. But it is not easy to do all these things. It is one of the things that highly differentiates Vertica from the rest of the pack. It is also why our customers, all of you continue to bet on us and see the value that we are delivering and we will continue to deliver. Here's a couple of examples of some of our customers who are powered by Vertica. It's the scale of data. It's the millisecond response times. Performance and scale have always been a huge part of what we have been about, not the only thing. I think the functionality all the capabilities that we add to the platform, the ease of use, the flexibility, obviously with the deployment. But if you look at some of the numbers they are under these customers on this slide. And I've shared a lot of different stories about these customers. Which, by the way, it still amaze me every time I talk to one and I get the updates, you can see the power and the difference that Vertica is making. Equally important, if you look at a lot of these customers, they are the epitome of being able to deploy Vertica in a lot of different environments. Many of the customers on this slide are not using Vertica just on-premise or just in the cloud. They're using it in a hybrid way. They're using it in multiple different clouds. And again, we've been with them on that journey throughout, which is what has made this product and frankly, our roadmap and our vision exactly what it is. It's been quite a journey. And that journey continues now with the Vertica 10 release. The Vertica 10 release is obviously a massive release for us. But if you look back, you can see that building on that native columnar architecture that started a long time ago, obviously, with the C-Store paper. We built it to leverage that commodity hardware, because it was an architecture that was never tightly integrated with any specific underlying infrastructure. I still remember hearing the initial pitch from Mike Stonebreaker, about the vision of Vertica as a software only solution and the importance of separating the company from hardware innovation. And at the time, Mike basically said to me, "there's so much R&D in innovation that's going to happen in hardware, we shouldn't bake hardware into our solution. We should do it in software, and we'll be able to take advantage of that hardware." And that is exactly what has happened. But one of the most recent innovations that we embraced with hardware is certainly that separation of compute and storage. As I said previously, the public cloud providers offered this next generation architecture, really to ensure that they can provide the customers exactly what they needed, more compute or more storage and charge for each, respectively. The separation of compute and storage, compute from storage is a major milestone in data center architectures. If you think about it, it's really not only a public cloud innovation, though. It fundamentally redefines the next generation data architecture for on-premise and for pretty much every way people are thinking about computing today. And that goes for software too. Object storage is an example of the cost effective means for storing data. And even more importantly, separating compute from storage for analytic workloads has a lot of advantages. Including the opportunity to manage much more dynamic, flexible workloads. And more importantly, truly isolate those workloads from others. And by the way, once you start having something that can truly isolate workloads, then you can have the conversations around autonomic computing, around setting up some nodes, some compute resources on the data that won't affect any of the other data to do some things on their own, maybe some self analytics, by the system, etc. A lot of things that many of you know we've already been exploring in terms of our own system data in the product. But it was May 2018, believe it or not, it seems like a long time ago where we first announced Eon Mode and I want to make something very clear, actually about Eon mode. It's a mode, it's a deployment option for Vertica customers. And I think this is another huge benefit that we don't talk about enough. But unlike a lot of vendors in the market who will dig you and charge you for every single add-on like hit-buy, you name it. You get this with the Vertica product. If you continue to pay support and maintenance, this comes with the upgrade. This comes as part of the new release. So any customer who owns or buys Vertica has the ability to set up either an Enterprise Mode or Eon Mode, which is a question I know that comes up sometimes. Our first announcement of Eon was obviously AWS customers, including the trade desk, AT&T. Most of whom will be speaking here later at the Virtual Big Data Conference. They saw a huge opportunity. Eon Mode, not only allowed Vertica to scale elastically with that specific compute and storage that was needed, but it really dramatically simplified database operations including things like workload balancing, node recovery, compute provisioning, etc. So one of the most popular functions is that ability to isolate the workloads and really allocate those resources without negatively affecting others. And even though traditional data warehouses, including Vertica Enterprise Mode have been able to do lots of different workload isolation, it's never been as strong as Eon Mode. Well, it certainly didn't take long for our customers to see that value across the board with Eon Mode. Not just up in the cloud, in partnership with one of our most valued partners and a platinum sponsor here. Joy mentioned at the beginning. We announced Vertica Eon Mode for Pure Storage FlashBlade in September 2019. And again, just to be clear, this is not a new product, it's one Vertica with yet more deployment options. With Pure Storage, Vertica in Eon mode is not limited in any way by variable cloud, network latency. The performance is actually amazing when you take the benefits of separate and compute from storage and you run it with a Pure environment on-premise. Vertica in Eon Mode has a super smart cache layer that we call the depot. It's a big part of our secret sauce around Eon mode. And combined with the power and performance of Pure's FlashBlade, Vertica became the industry's first advanced analytics platform that actually separates compute and storage for on-premises data centers. Something that a lot of our customers are already benefiting from, and we're super excited about it. But as I said, this is a journey. We don't stop, we're not going to stop. Our customers need the flexibility of multiple public clouds. So today with Vertica 10, we're super proud and excited to announce support for Vertica in Eon Mode on Google Cloud. This gives our customers the ability to use their Vertica licenses on Amazon AWS, on-premise with Pure Storage and on Google Cloud. Now, we were talking about HDFS and a lot of our customers who have invested quite a bit in HDFS as a place, especially to store data have been pushing us to support Eon Mode with HDFS. So as part of Vertica 10, we are also announcing support for Vertica in Eon Mode using HDFS as the communal storage. Vertica's own Roth format data can be stored in HDFS, and actually the full functionality of Vertica is complete analytics, geospatial pattern matching, time series, machine learning, everything that we have in there can be applied to this data. And on the same HDFS nodes, Vertica can actually also analyze data in ORC or Parquet format, using External tables. We can also execute joins between the Roth data the External table holds, which powers a much more comprehensive view. So again, it's that flexibility to be able to support our customers, wherever they need us to support them on whatever platform, they have. Vertica 10 gives us a lot more ways that we can deploy Eon Mode in various environments for our customers. It allows them to take advantage of Vertica in Eon Mode and the power that it brings with that separation, with that workload isolation, to whichever platform they are most comfortable with. Now, there's a lot that has come in Vertica 10. I'm definitely not going to be able to cover everything. But we also introduced complex types as an example. And complex data types fit very well into Eon as well in this separation. They significantly reduce the data pipeline, the cost of moving data between those, a much better support for unstructured data, which a lot of our customers have mixed with structured data, of course, and they leverage a lot of columnar execution that Vertica provides. So you get complex data types in Vertica now, a lot more data, stronger performance. It goes great with the announcement that we made with the broader Eon Mode. Let's talk a little bit more about machine learning. We've been actually doing work in and around machine learning with various extra regressions and a whole bunch of other algorithms for several years. We saw the huge advantage that MPP offered, not just as a sequel engine as a database, but for ML as well. Didn't take as long to realize that there's a lot more to operationalizing machine learning than just those algorithms. It's data preparation, it's that model trade training. It's the scoring, the shaping, the evaluation. That is so much of what machine learning and frankly, data science is about. You do know, everybody always wants to jump to the sexy algorithm and we handle those tasks very, very well. It makes Vertica a terrific platform to do that. A lot of work in data science and machine learning is done in other tools. I had mentioned that there's just so many tools out there. We want people to be able to take advantage of all that. We never believed we were going to be the best algorithm company or come up with the best models for people to use. So with Vertica 10, we support PMML. We can import now and export PMML models. It's a huge step for us around that operationalizing machine learning projects for our customers. Allowing the models to get built outside of Vertica yet be imported in and then applying to that full scale of data with all the performance that you would expect from Vertica. We also are more tightly integrating with Python. As many of you know, we've been doing a lot of open source projects with the community driven by many of our customers, like Uber. And so now with Python we've integrated with TensorFlow, allowing data scientists to build models in their preferred language, to take advantage of TensorFlow. But again, to store and deploy those models at scale with Vertica. I think both these announcements are proof of our big bet number three, and really our commitment to supporting innovation throughout the community by operationalizing ML with that accuracy, performance and scale of Vertica for our customers. Again, there's a lot of steps when it comes to the workflow of machine learning. These are some of them that you can see on the slide, and it's definitely not linear either. We see this as a circle. And companies that do it, well just continue to learn, they continue to rescore, they continue to redeploy and they want to operationalize all that within a single platform that can take advantage of all those capabilities. And that is the platform, with a very robust ecosystem that Vertica has always been committed to as an organization and will continue to be. This graphic, many of you have seen it evolve over the years. Frankly, if we put everything and everyone on here wouldn't fit on a slide. But it will absolutely continue to evolve and grow as we support our customers, where they need the support most. So, again, being able to deploy everywhere, being able to take advantage of Vertica, not just as a business analyst or a business user, but as a data scientists or as an operational or BI person. We want Vertica to be leveraged and used by the broader organization. So I think it's fair to say and I encourage everybody to learn more about Vertica 10, because I'm just highlighting some of the bigger aspects of it. But we talked about those three market trends. The need to unify the silos, the need for hybrid multiple cloud deployment options, the need to operationalize business critical machine learning projects. Vertica 10 has absolutely delivered on those. But again, we are not going to stop. It is our job not to, and this is how Team Vertica thrives. I always joke that the next release is the best release. And, of course, even after Vertica 10, that is also true, although Vertica 10 is pretty awesome. But, you know, from the first line of code, we've always been focused on performance and scale, right. And like any really strong data platform, the execution engine, the optimizer and the execution engine are the two core pieces of that. Beyond Vertica 10, some of the big things that we're already working on, next generation execution engine. We're already actually seeing incredible early performance from this. And this is just one example, of how important it is for an organization like Vertica to constantly go back and re-innovate. Every single release, we do the sit ups and crunches, our performance and scale. How do we improve? And there's so many parts of the core server, there's so many parts of our broader ecosystem. We are constantly looking at coverages of how we can go back to all the code lines that we have, and make them better in the current environment. And it's not an easy thing to do when you're doing that, and you're also expanding in the environment that we are expanding into to take advantage of the different deployments, which is a great segue to this slide. Because if you think about today, we're obviously already available with Eon Mode and Amazon, AWS and Pure and actually MinIO as well. As I talked about in Vertica 10 we're adding Google and HDFS. And coming next, obviously, Microsoft Azure, Alibaba cloud. So being able to expand into more of these environments is really important for the Vertica team and how we go forward. And it's not just running in these clouds, for us, we want it to be a SaaS like experience in all these clouds. We want you to be able to deploy Vertica in 15 minutes or less on these clouds. You can also consume Vertica, in a lot of different ways, on these clouds. As an example, in Amazon Vertica by the Hour. So for us, it's not just about running, it's about taking advantage of the ecosystems that all these cloud providers offer, and really optimizing the Vertica experience as part of them. Optimization, around automation, around self service capabilities, extending our management console, we now have products that like the Vertica Advisor Tool that our Customer Success Team has created to actually use our own smarts in Vertica. To take data from customers that give it to us and help them tune automatically their environment. You can imagine that we're taking that to the next level, in a lot of different endeavors that we're doing around how Vertica as a product can actually be smarter because we all know that simplicity is key. There just aren't enough people in the world who are good at managing data and taking it to the next level. And of course, other things that we all hear about, whether it's Kubernetes and containerization. You can imagine that that probably works very well with the Eon Mode and separating compute and storage. But innovation happens everywhere. We innovate around our community documentation. Many of you have taken advantage of the Vertica Academy. The numbers there are through the roof in terms of the number of people coming in and certifying on it. So there's a lot of things that are within the core products. There's a lot of activity and action beyond the core products that we're taking advantage of. And let's not forget why we're here, right? It's easy to talk about a platform, a data platform, it's easy to jump into all the functionality, the analytics, the flexibility, how we can offer it. But at the end of the day, somebody, a person, she's got to take advantage of this data, she's got to be able to take this data and use this information to make a critical business decision. And that doesn't happen unless we explore lots of different and frankly, new ways to get that predictive analytics UI and interface beyond just the standard BI tools in front of her at the right time. And so there's a lot of activity, I'll tease you with that going on in this organization right now about how we can do that and deliver that for our customers. We're in a great position to be able to see exactly how this data is consumed and used and start with this core platform that we have to go out. Look, I know, the plan wasn't to do this as a virtual BDC. But I really appreciate you tuning in. Really appreciate your support. I think if there's any silver lining to us, maybe not being able to do this in person, it's the fact that the reach has actually gone significantly higher than what we would have been able to do in person in Boston. We're certainly looking forward to doing a Big Data Conference in the future. But if I could leave you with anything, know this, since that first release for Vertica, and our very first customers, we have been very consistent. We respect all the innovation around us, whether it's open source or not. We understand the market trends. We embrace those new ideas and technologies and for us true north, and the most important thing is what does our customer need to do? What problem are they trying to solve? And how do we use the advantages that we have without disrupting our customers? But knowing that you depend on us to deliver that unified analytics strategy, it will deliver that performance of scale, not only today, but tomorrow and for years to come. We've added a lot of great features to Vertica. I think we've said no to a lot of things, frankly, that we just knew we wouldn't be the best company to deliver. When we say we're going to do things we do them. Vertica 10 is a perfect example of so many of those things that we from you, our customers have heard loud and clear, and we have delivered. I am incredibly proud of this team across the board. I think the culture of Vertica, a customer first culture, jumping in to help our customers win no matter what is also something that sets us massively apart. I hear horror stories about support experiences with other organizations. And people always seem to be amazed at Team Vertica's willingness to jump in or their aptitude for certain technical capabilities or understanding the business. And I think sometimes we take that for granted. But that is the team that we have as Team Vertica. We are incredibly excited about Vertica 10. I think you're going to love the Virtual Big Data Conference this year. I encourage you to tune in. Maybe one other benefit is I know some people were worried about not being able to see different sessions because they were going to overlap with each other well now, even if you can't do it live, you'll be able to do those sessions on demand. Please enjoy the Vertica Big Data Conference here in 2020. Please you and your families and your co-workers be safe during these times. I know we will get through it. And analytics is probably going to help with a lot of that and we already know it is helping in many different ways. So believe in the data, believe in data's ability to change the world for the better. And thank you for your time. And with that, I am delighted to now introduce Micro Focus CEO Stephen Murdoch to the Vertica Big Data Virtual Conference. Thank you Stephen. >> Stephen: Hi, everyone, my name is Stephen Murdoch. I have the pleasure and privilege of being the Chief Executive Officer here at Micro Focus. Please let me add my welcome to the Big Data Conference. And also my thanks for your support, as we've had to pivot to this being virtual rather than a physical conference. Its amazing how quickly we all reset to a new normal. I certainly didn't expect to be addressing you from my study. Vertica is an incredibly important part of Micro Focus family. Is key to our goal of trying to enable and help customers become much more data driven across all of their IT operations. Vertica 10 is a huge step forward, we believe. It allows for multi-cloud innovation, genuinely hybrid deployments, begin to leverage machine learning properly in the enterprise, and also allows the opportunity to unify currently siloed lakes of information. We operate in a very noisy, very competitive market, and there are people, who are in that market who can do some of those things. The reason we are so excited about Vertica is we genuinely believe that we are the best at doing all of those things. And that's why we've announced publicly, you're under executing internally, incremental investment into Vertica. That investments targeted at accelerating the roadmaps that already exist. And getting that innovation into your hands faster. This idea is speed is key. It's not a question of if companies have to become data driven organizations, it's a question of when. So that speed now is really important. And that's why we believe that the Big Data Conference gives a great opportunity for you to accelerate your own plans. You will have the opportunity to talk to some of our best architects, some of the best development brains that we have. But more importantly, you'll also get to hear from some of our phenomenal Roth Data customers. You'll hear from Uber, from the Trade Desk, from Philips, and from AT&T, as well as many many others. And just hearing how those customers are using the power of Vertica to accelerate their own, I think is the highlight. And I encourage you to use this opportunity to its full. Let me close by, again saying thank you, we genuinely hope that you get as much from this virtual conference as you could have from a physical conference. And we look forward to your engagement, and we look forward to hearing your feedback. With that, thank you very much. >> Joy: Thank you so much, Stephen, for joining us for the Vertica Big Data Conference. Your support and enthusiasm for Vertica is so clear, and it makes a big difference. Now, I'm delighted to introduce Amy Fowler, the VP of Strategy and Solutions for FlashBlade at Pure Storage, who was one of our BDC Platinum Sponsors, and one of our most valued partners. It was a proud moment for me, when we announced Vertica in Eon mode for Pure Storage FlashBlade and we became the first analytics data warehouse that separates compute from storage for on-premise data centers. Thank you so much, Amy, for joining us. Let's get started. >> Amy: Well, thank you, Joy so much for having us. And thank you all for joining us today, virtually, as we may all be. So, as we just heard from Colin Mahony, there are some really interesting trends that are happening right now in the big data analytics market. From the end of the Hadoop hype cycle, to the new cloud reality, and even the opportunity to help the many data science and machine learning projects move from labs to production. So let's talk about these trends in the context of infrastructure. And in particular, look at why a modern storage platform is relevant as organizations take on the challenges and opportunities associated with these trends. The answer is the Hadoop hype cycles left a lot of data in HDFS data lakes, or reservoirs or swamps depending upon the level of the data hygiene. But without the ability to get the value that was promised from Hadoop as a platform rather than a distributed file store. And when we combine that data with the massive volume of data in Cloud Object Storage, we find ourselves with a lot of data and a lot of silos, but without a way to unify that data and find value in it. Now when you look at the infrastructure data lakes are traditionally built on, it is often direct attached storage or data. The approach that Hadoop took when it entered the market was primarily bound by the limits of networking and storage technologies. One gig ethernet and slower spinning disk. But today, those barriers do not exist. And all FlashStorage has fundamentally transformed how data is accessed, managed and leveraged. The need for local data storage for significant volumes of data has been largely mitigated by the performance increases afforded by all Flash. At the same time, organizations can achieve superior economies of scale with that segregation of compute and storage. With compute and storage, you don't always scale in lockstep. Would you want to add an engine to the train every time you add another boxcar? Probably not. But from a Pure Storage perspective, FlashBlade is uniquely architected to allow customers to achieve better resource utilization for compute and storage, while at the same time, reducing complexity that has arisen from the siloed nature of the original big data solutions. The second and equally important recent trend we see is something I'll call cloud reality. The public clouds made a lot of promises and some of those promises were delivered. But cloud economics, especially usage based and elastic scaling, without the control that many companies need to manage the financial impact is causing a lot of issues. In addition, the risk of vendor lock-in from data egress, charges, to integrated software stacks that can't be moved or deployed on-premise is causing a lot of organizations to back off the all the way non-cloud strategy, and move toward hybrid deployments. Which is kind of funny in a way because it wasn't that long ago that there was a lot of talk about no more data centers. And for example, one large retailer, I won't name them, but I'll admit they are my favorites. They several years ago told us they were completely done with on-prem storage infrastructure, because they were going 100% to the cloud. But they just deployed FlashBlade for their data pipelines, because they need predictable performance at scale. And the all cloud TCO just didn't add up. Now, that being said, well, there are certainly challenges with the public cloud. It has also brought some things to the table that we see most organizations wanting. First of all, in a lot of cases applications have been built to leverage object storage platforms like S3. So they need that object protocol, but they may also need it to be fast. And the said object may be oxymoron only a few years ago, and this is an area of the market where Pure and FlashBlade have really taken a leadership position. Second, regardless of where the data is physically stored, organizations want the best elements of a cloud experience. And for us, that means two main things. Number one is simplicity and ease of use. If you need a bunch of storage experts to run the system, that should be considered a bug. The other big one is the consumption model. The ability to pay for what you need when you need it, and seamlessly grow your environment over time totally nondestructively. This is actually pretty huge and something that a lot of vendors try to solve for with finance programs. But no finance program can address the pain of a forklift upgrade, when you need to move to next gen hardware. To scale nondestructively over long periods of time, five to 10 years plus is a crucial architectural decisions need to be made at the outset. Plus, you need the ability to pay as you use it. And we offer something for FlashBlade called Pure as a Service, which delivers exactly that. The third cloud characteristic that many organizations want is the option for hybrid. Even if that is just a DR site in the cloud. In our case, that means supporting appplication of S3, at the AWS. And the final trend, which to me represents the biggest opportunity for all of us, is the need to help the many data science and machine learning projects move from labs to production. This means bringing all the machine learning functions and model training to the data, rather than moving samples or segments of data to separate platforms. As we all know, machine learning needs a ton of data for accuracy. And there is just too much data to retrieve from the cloud for every training job. At the same time, predictive analytics without accuracy is not going to deliver the business advantage that everyone is seeking. You can kind of visualize data analytics as it is traditionally deployed as being on a continuum. With that thing, we've been doing the longest, data warehousing on one end, and AI on the other end. But the way this manifests in most environments is a series of silos that get built up. So data is duplicated across all kinds of bespoke analytics and AI, environments and infrastructure. This creates an expensive and complex environment. So historically, there was no other way to do it because some level of performance is always table stakes. And each of these parts of the data pipeline has a different workload profile. A single platform to deliver on the multi dimensional performances, diverse set of applications required, that didn't exist three years ago. And that's why the application vendors pointed you towards bespoke things like DAS environments that we talked about earlier. And the fact that better options exists today is why we're seeing them move towards supporting this disaggregation of compute and storage. And when it comes to a platform that is a better option, one with a modern architecture that can address the diverse performance requirements of this continuum, and allow organizations to bring a model to the data instead of creating separate silos. That's exactly what FlashBlade is built for. Small files, large files, high throughput, low latency and scale to petabytes in a single namespace. And this is importantly a single rapid space is what we're focused on delivering for our customers. At Pure, we talk about it in the context of modern data experience because at the end of the day, that's what it's really all about. The experience for your teams in your organization. And together Pure Storage and Vertica have delivered that experience to a wide range of customers. From a SaaS analytics company, which uses Vertica on FlashBlade to authenticate the quality of digital media in real time, to a multinational car company, which uses Vertica on FlashBlade to make thousands of decisions per second for autonomous cars, or a healthcare organization, which uses Vertica on FlashBlade to enable healthcare providers to make real time decisions that impact lives. And I'm sure you're all looking forward to hearing from John Yavanovich from AT&T. To hear how he's been doing this with Vertica and FlashBlade as well. He's coming up soon. We have been really excited to build this partnership with Vertica. And we're proud to provide the only on-premise storage platform validated with Vertica Eon Mode. And deliver this modern data experience to our customers together. Thank you all so much for joining us today. >> Joy: Amy, thank you so much for your time and your insights. Modern infrastructure is key to modern analytics, especially as organizations leverage next generation data center architectures, and object storage for their on-premise data centers. Now, I'm delighted to introduce our last speaker in our Vertica Big Data Conference Keynote, John Yovanovich, Director of IT for AT&T. Vertica is so proud to serve AT&T, and especially proud of the harmonious impact we are having in partnership with Pure Storage. John, welcome to the Virtual Vertica BDC. >> John: Thank you joy. It's a pleasure to be here. And I'm excited to go through this presentation today. And in a unique fashion today 'cause as I was thinking through how I wanted to present the partnership that we have formed together between Pure Storage, Vertica and AT&T, I want to emphasize how well we all work together and how these three components have really driven home, my desire for a harmonious to use your word relationship. So, I'm going to move forward here and with. So here, what I'm going to do the theme of today's presentation is the Pure Vertica Symphony live at AT&T. And if anybody is a Westworld fan, you can appreciate the sheet music on the right hand side. What we're going to what I'm going to highlight here is in a musical fashion, is how we at AT&T leverage these technologies to save money to deliver a more efficient platform, and to actually just to make our customers happier overall. So as we look back, and back as early as just maybe a few years ago here at AT&T, I realized that we had many musicians to help the company. Or maybe you might want to call them data scientists, or data analysts. For the theme we'll stay with musicians. None of them were singing or playing from the same hymn book or sheet music. And so what we had was many organizations chasing a similar dream, but not exactly the same dream. And, best way to describe that is and I think with a lot of people this might resonate in your organizations. How many organizations are chasing a customer 360 view in your company? Well, I can tell you that I have at least four in my company. And I'm sure there are many that I don't know of. That is our problem because what we see is a repetitive sourcing of data. We see a repetitive copying of data. And there's just so much money to be spent. This is where I asked Pure Storage and Vertica to help me solve that problem with their technologies. What I also noticed was that there was no coordination between these departments. In fact, if you look here, nobody really wants to play with finance. Sales, marketing and care, sure that you all copied each other's data. But they actually didn't communicate with each other as they were copying the data. So the data became replicated and out of sync. This is a challenge throughout, not just my company, but all companies across the world. And that is, the more we replicate the data, the more problems we have at chasing or conquering the goal of single version of truth. In fact, I kid that I think that AT&T, we actually have adopted the multiple versions of truth, techno theory, which is not where we want to be, but this is where we are. But we are conquering that with the synergies between Pure Storage and Vertica. This is what it leaves us with. And this is where we are challenged and that if each one of our siloed business units had their own stories, their own dedicated stories, and some of them had more money than others so they bought more storage. Some of them anticipating storing more data, and then they really did. Others are running out of space, but can't put anymore because their bodies aren't been replenished. So if you look at it from this side view here, we have a limited amount of compute or fixed compute dedicated to each one of these silos. And that's because of the, wanting to own your own. And the other part is that you are limited or wasting space, depending on where you are in the organization. So there were the synergies aren't just about the data, but actually the compute and the storage. And I wanted to tackle that challenge as well. So I was tackling the data. I was tackling the storage, and I was tackling the compute all at the same time. So my ask across the company was can we just please play together okay. And to do that, I knew that I wasn't going to tackle this by getting everybody in the same room and getting them to agree that we needed one account table, because they will argue about whose account table is the best account table. But I knew that if I brought the account tables together, they would soon see that they had so much redundancy that I can now start retiring data sources. I also knew that if I brought all the compute together, that they would all be happy. But I didn't want them to tackle across tackle each other. And in fact that was one of the things that all business units really enjoy. Is they enjoy the silo of having their own compute, and more or less being able to control their own destiny. Well, Vertica's subclustering allows just that. And this is exactly what I was hoping for, and I'm glad they've brought through. And finally, how did I solve the problem of the single account table? Well when you don't have dedicated storage, and you can separate compute and storage as Vertica in Eon Mode does. And we store the data on FlashBlades, which you see on the left and right hand side, of our container, which I can describe in a moment. Okay, so what we have here, is we have a container full of compute with all the Vertica nodes sitting in the middle. Two loader, we'll call them loader subclusters, sitting on the sides, which are dedicated to just putting data onto the FlashBlades, which is sitting on both ends of the container. Now today, I have two dedicated storage or common dedicated might not be the right word, but two storage racks one on the left one on the right. And I treat them as separate storage racks. They could be one, but i created them separately for disaster recovery purposes, lashing work in case that rack were to go down. But that being said, there's no reason why I'm probably going to add a couple of them here in the future. So I can just have a, say five to 10, petabyte storage, setup, and I'll have my DR in another 'cause the DR shouldn't be in the same container. Okay, but I'll DR outside of this container. So I got them all together, I leveraged subclustering, I leveraged separate and compute. I was able to convince many of my clients that they didn't need their own account table, that they were better off having one. I eliminated, I reduced latency, I reduced our ticketing I reduce our data quality issues AKA ticketing okay. I was able to expand. What is this? As work. I was able to leverage elasticity within this cluster. As you can see, there are racks and racks of compute. We set up what we'll call the fixed capacity that each of the business units needed. And then I'm able to ramp up and release the compute that's necessary for each one of my clients based on their workloads throughout the day. And so while they compute to the right before you see that the instruments have already like, more or less, dedicated themselves towards all those are free for anybody to use. So in essence, what I have, is I have a concert hall with a lot of seats available. So if I want to run a 10 chair Symphony or 80, chairs, Symphony, I'm able to do that. And all the while, I can also do the same with my loader nodes. I can expand my loader nodes, to actually have their own Symphony or write all to themselves and not compete with any other workloads of the other clusters. What does that change for our organization? Well, it really changes the way our database administrators actually do their jobs. This has been a big transformation for them. They have actually become data conductors. Maybe you might even call them composers, which is interesting, because what I've asked them to do is morph into less technology and more workload analysis. And in doing so we're able to write auto-detect scripts, that watch the queues, watch the workloads so that we can help ramp up and trim down the cluster and subclusters as necessary. There has been an exciting transformation for our DBAs, who I need to now classify as something maybe like DCAs. I don't know, I have to work with HR on that. But I think it's an exciting future for their careers. And if we bring it all together, If we bring it all together, and then our clusters, start looking like this. Where everything is moving in harmonious, we have lots of seats open for extra musicians. And we are able to emulate a cloud experience on-prem. And so, I want you to sit back and enjoy the Pure Vertica Symphony live at AT&T. (soft music) >> Joy: Thank you so much, John, for an informative and very creative look at the benefits that AT&T is getting from its Pure Vertica symphony. I do really like the idea of engaging HR to change the title to Data Conductor. That's fantastic. I've always believed that music brings people together. And now it's clear that analytics at AT&T is part of that musical advantage. So, now it's time for a short break. And we'll be back for our breakout sessions, beginning at 12 pm Eastern Daylight Time. We have some really exciting sessions planned later today. And then again, as you can see on Wednesday. Now because all of you are already logged in and listening to this keynote, you already know the steps to continue to participate in the sessions that are listed here and on the previous slide. In addition, everyone received an email yesterday, today, and you'll get another one tomorrow, outlining the simple steps to register, login and choose your session. If you have any questions, check out the emails or go to www.vertica.com/bdc2020 for the logistics information. There are a lot of choices and that's always a good thing. Don't worry if you want to attend one or more or can't listen to these live sessions due to your timezone. All the sessions, including the Q&A sections will be available on demand and everyone will have access to the recordings as well as even more pre-recorded sessions that we'll post to the BDC website. Now I do want to leave you with two other important sites. First, our Vertica Academy. Vertica Academy is available to everyone. And there's a variety of very technical, self-paced, on-demand training, virtual instructor-led workshops, and Vertica Essentials Certification. And it's all free. Because we believe that Vertica expertise, helps everyone accelerate their Vertica projects and the advantage that those projects deliver. Now, if you have questions or want to engage with our Vertica engineering team now, we're waiting for you on the Vertica forum. We'll answer any questions or discuss any ideas that you might have. Thank you again for joining the Vertica Big Data Conference Keynote Session. Enjoy the rest of the BDC because there's a lot more to come