So, click a button, move the data. If you want it to be a replica, click a button and say "replicate." If you want to just move it, just click a button and move it. It's literally that easy. >> And so the customers can choose where to put the data.
>> Yes. >> Can they do a public version of this, or only private? >> Both, it connects to public as well https://casinoslots.sg/online-roulette. >> Okay, so that was Jenny's mention, okay cool. What's the most exciting thing for you this week going on in your world? Obviously, center of the value proposition, and Jenny used your lines so I'm sure you fed her some good sound bytes there, because she was basically taking your pitch as the headline for the keynote. Is that the highlight, or is it customer activity? >> I think the exciting thing, and Jenny did talk about it, is connecting data to AI. I'd say many clients have kind of thought of those as two different topics. We do that in three ways. We say common machine learning fabric. You can build a model in Watson, you can deploy it where your enterprise data is or vice versa. We do that with the metadata. You create business or technical metadata on-premise, you can push that to Watson or vice versa. And like we just talked about, we make the data movement incredibly easy. So we're uniting these two worlds of data and AI that have tended to be different parts of an organization in many clients. We're uniting that, I think that's pretty interesting. >> All right, so final question, I've got to ask the tough one, which is, okay, Rob I love it, but I'm really not paying attention to the data because I've got my hands full in my IT transformation and we're making critical decisions on cloud globally, I've got multiple regions to deal with, I got different issues outside in each digital nation, but I'm going to get the data after. What's in it for me, your whole pitch? I'm dealing with cloud right now, so why should I be cross-connecting with the cloud decision and the cloud conversations that relate to the benefit of what you're doing? >> If you're not paying attention to the data, you're not going to be around. So your cloud decisions are kind of worthless, because you're not going to be around if you're not paying attention to the data. >> So I can make a bad cloud decision if I don't factor in what? >> I believe you have to think about your data strategy. Look, every organization is going to be multi-cloud, but you have to have a single data strategy regardless of what your cloud strategy is. You've got to think about all those building blocks I talked about. Manage data, collect data, govern data, analyze data. That has to be one strategy regardless of cloud. If you're not thinking about that, you're in trouble. >> Or making sure that I have Kubernetes? Is that a good decision? >> That is a great decision. >> (laughs) >> Makes it really easy, seamless to deploy applications, to deploy data, to move it around clouds. Makes it really easy. >> And what's the business model for containers? Kind of shifts to being a commodity? >> I think over time, yes, but there's so much to do around containers because containers, again, go back to the analogy. It's just the crate. >> John: Makes things easy. >> It's not the cargo, it's not the ship. It's just the crate, it's one piece. >> Yeah, and there's no, a lot of choice there, too. Clients can do whatever they want. >> Yeah. All right, we love Kubernetes. We'll be at KubeCon in Copenhagen next month, so keep a lookout there for us. This is Rob Thomas, here inside the Cube, here at IBM Think, breaking down all the action in the data science world, data world. It's the center of the value proposition. Main story here at IBM Think is data at the center of the value proposition for the modern enterprise. I'm John Furrier inside the Cube. We'll be back with more after this short break. (light electronic music)
0 Comments
How does that fit for data? I'm just putting a container on data sets? Who's addressing the envelope of that container?
How is that addressable? I mean, how does it work? >> Let me give you an analogy. So you go back to the year 1955. There is no standards in any shipping port around the world. Everybody is literally building their own containers, building their own ships, building their own trucks. It's incredibly expensive and takes forever to get cargo to move from one place to the next. 1956, a guy named Malcom McClean, he invents the first intermodal shipping container, patents it. It becomes the standard. So now, every port, every container looks identical. What's the benefit? Sure, it made more flexibility. Saved lot of money, 90% of the cost came out of shipping a container. But the biggest thing is it changed commerce. So, you look at GDP at that time, it took off. All because of the standardization around a form factor that made it accessible to everybody. Now, let's put that in the IT world. We got containers for the application world. Made it much easier to deploy, a standard, again. >> Yeah, and program around. >> More cost-effective, more-- Yep, exactly. What's the cargo in IT? It's data. Data is the cargo, that's what's sitting inside the container. Now you have to say, how do we actually take the same concepts that we did for applications, make that available for data so that my data can fit anywhere? That's what we're doing. >> How does that work and what's the impact to the customer? >> Is it IBM software that you're doing? Is it Kubernetes open source software? Just tie that together for me. So IBM Cloud Private is our Kubernetes distribution, with some different pieces we put on it. When you add the Cloud Private for Data, it's got a Spark Engine, like everything we do it's based on open source to start with. And then we have an experience for a data scientist, an experience for a data analyst. It's your view to your enterprise data. You'll love the UI when you see it. First, above the fold, all my machine learning models in the organization, what's working, what's not working. Below the fold, what's my data? Structured or unstructured? Sensitive, non-sensitive? I click it on, I can see all of my data. Hadoop, Cloud-A, Cloud-B, Cloud-C, on-premise system. It's get a view to all of your data. >> So is the purpose to move the data around? >> No, the purpose is actually the exact opposite. Leave the data in place, but be able to treat it as a single data environment. We're doing a lot of work with Federation, our SQL technology which historically, as we all know, Federation hasn't really performed. We have it performing. >> Okay, so I'm just, in the use case in my head, so I store the data on my private, secure, comfortable, feeling good about it, but I have a public cloud app. How does that work? Is it a replica of the data? Is it just the container that makes it addressable? How does that move across? Without those things, you just have a shiny object and not necessarily an outcome. That's why these building blocks are fundamental. And the clients, they get to this point, and they're the ones who try to jump to the shiny object and they don't have the data to support that. >> And then you've got companies going on digital transformation, which is basically saying all their data legacy, trying to modernize it. The modern companies like Uber, and we saw the first fatality of an Uber car this week, again, points out the reality that realtime is realtime, and the importance of having data, whether it's sensing data.
We're not, it's coming there, you can start to see it happening. Realtime data is key. That means data mobility is critical, and you mentioned private, public. Storing the data and moving the data around, having data intelligence, is the most important thing. Realtime data in motion, intelligence, you know, where are we? Is that a setback with the Uber incident? Is it a step forward, is it learning? What's your view of the data quality of movement in realtime? >> I think data ingestion is one of the least talked about topics that is one of the most important. With IBM Cloud Private for data, we can ingest 250 billion events a day. Let me give you some context for that. 2016, the entire credit card industry, everywhere in the world, did 250 billion transactions. So what credit cards do in a year, we can do in a day. Biggest stock trading day ever on the New York Stock Exchange, what got done in that entire day, we can do in the first 40 minutes of trading. But that value there is, how fast can you bring data in to be analyzed, and can you do a decent bit of that pre-processing, or analytics, on the way in? That's how you start to solve some of the problems that you're describing, because it's instant >> John: Yeah. >> And it's unsurpassed amounts of data. >> So ingestion's a key part of the value chain, if you will, on data management. The new kind of data management. Ingesting it, understanding context, then is that where AI kicks in? Where does the AI kick in? Because the ingestion speaks to the information architecture, IA. >> Rob: Yes. >> Now I got to put AI on top of that data, so is the data different? Talk about the dynamic between, okay I'm ingesting data for the sake of ingesting, where does the AI connect? >> So you got the data, yep. So you go the data, AI starts where you're saying, all right, now we want to automate this. We're going to build models, we're going to use the data that we've got in here to train those models. As we get more data, the models are going to get better. Now we're going to connect it to how humans want to interact. Maybe it's natural language processing, maybe it's visualizing data. That's the whole lineage of how somebody gets toward this AI idea. >> What are some of the conversations you're having with customers, and how have they changed? And give some color, I mean, only a few years ago we're talking about data lakes. >> Right. >> Okay, what is the conversation now, and give some context of how far that conversation has gone down the road toward advancement. >> I think we're going from data lakes to an idea of a fluid data layer, which is all your data assets managed as a single system, even if they sit in different architectures. Because there's no one, we all know this. We've been around this industry forever. There's no one way to support or manage data that's going to support every use case. So this idea of a fluid data layer becomes critical for every organization. That's one big change. Other big change is containers. What we're doing with Cloud Private for Data is based on Kubernetes, that's how people want to consume applications, but nobody's really solved that for data. I think we're solving that for data. >> Let's dig into that. It was one of my topics I wanted to drill down on. Containers have been great for moving workloads around, certainly Kubernetes has been a great orchestration tool. |
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
December 2018
Categories |