Alexia, China, Chino discussed on O'Reilly Data Show

Automatic TRANSCRIPT

Very exciting. They do model training they do risk management. They do more training for tweeting. They do I don't highs meant recommendation like all kinds of different applications super exciting. So as best as you can tell h why a lot of the popular open source machine. Learning ivories being used in conjunction with Alexia, right? Yes. Several weeks ago. There was a meet up in Shanghai. And there's a company called a chino. The are. A public company they're in China, and they have like a hundred people a team internally and they run their motion learning workloads on top of Oxy's on top of their object storage. And there are so many different off remarks. They're running on top of us and very exciting to see that and actually lost. We came back from New York and with many users in that region on using machine on top Las well, and many of these use cases directly brain like a business value or or more revenue to those companies, which is very exciting. So you mentioned earlier when you were describing Alexia debt one of the things that solves is you have many many different your might be working in an enterprise at many, different storage systems and Alexia sits in the middle between this this many different stories systems and your many different. Analytic libraries. But you you also just mentioned that in China. There's this company that has lost between the object store and and machine learning tool. So so one question is aren't the object store fast, distaste like s free and Seth. So why do I need a why do I need a layer in between? So there are two perspective. Two angles to answer. This question. One on goal is that one key reason people use object store, it is cheap and her gigabytes or Protais is cheaper than other solutions in a market and many people use of just or actually as Akobo storage and Palmas is not as good as what people wanted. And from that perspective by putting on Luxy on top of that. That's the late. I'm improved the performance from our caching functionality and on top of that many cases machine running elaborated are not directed talk with objects, doors and. The we Alexia also sir us said translation layer from that angle. That's one perspective. And the other perspective, he's kind of bigger picture by Luke Kuhn, a bigger picture all becomes wny's all the nations in the end of the day. They wanna build their data infrastructure. They the platform to serve their existing daydream applications like a militates Mercer money is such but moving forward there will be more. They drew applications in created by Hiram years innovators our industry when that happens like you need to have this great day the platform to serve existing workloads, and you're calls a contract with a listing storage and the new storage and the in that world this new architecture, which we call data ecosystem two point, oh with Alexia, this virtual issued it falls system layer in the middle. Just makes a lot of sons from architecture. Perspective and by actually based on our conversation with some top architects in the industry like people believe is a beautiful architecture very elegant and also future proven us. Well, and this is a bigger picture aside to make sure is enterprise, your company is successful. You need to be able to build the architecture or this platform for today in the morning points today for the future. So that's two angles. Why able you.

Coming up next