AWS reInvent 2019 - Analytics, ML, and Kubernetes
Transcript
>> Hiya doin'? This is Damani Corbin sitting in for Keith Townsend on The CTO Advisor. This is your CTO Dose coming to you live from Las Vegas. We're at AWS re:Invent 2019. I have the pleasure of sitting down with Kevin from GoodData, and Isi from Accenture. Kevin, you want to introduce yourself? >> Good afternoon, I'm Kevin young with GoodData. I'm the Global Director for Sales Engineering for GoodData. >> Perfect. Isi? >> I'm an AWS Cloud Supervision Associate Manager at the Accenture AWS Business Group.
>> Perfect. Today we'll be speaking about Kubernetes, we'll be speaking about the announcement that Andy Jassy made yesterday during his keynote. I'll touch a little bit about machine learning, and artificial intelligence. Thank you for listening in. >> Isi do you want to introduce yourself, and kind of your role? >> Yeah, my name is Isi Idemudia. I'm with Accenture. I'm focused on the AWS practice, so the Accenture AWS Business Group, and I'm a Solutions Architect at Accenture, and we build, design, manage, give best practices for AWS deployment.
>> Perfect. Kevin? >> I'm Kevin Young, I'm the Global Director for GoodData for sales engineering. >> Perfect, perfect. So, we're here at Amazon AWS re:Invent 2019. Want to try and just get your take on Andy Jassy's keynote yesterday. What was anything that excited you? Were there any things that you're excited to bring back to your customers? Yeah, I think the overall, it was great. I mean three hours of different, you know, features, and services. But what really was exciting for me was the machine learning part of things, and it was especially Sagemaker.
I've watched Sagemaker the last few years, and I've seen how the product has evolved. And, yeah, right now for Sagemaker, we have an IDE called the Sagemaker Studio, where developers can build machine-learning models from start to finish, and especially the Kubernetes part of things. Now, there's a Sagemaker Operator that runs on Kubernetes. So, it's fantastic to see how the product has matured. >> I'm very excited about the Kubernetes aspect of it, and we're going to double click on that a little bit.
Kevin, what are you excited about in this conference? >> I think there's been a lot of interesting conversations with customers, and prospects, who are all looking to move to the cloud, especially multi-cloud environments, and how to deploy that. And, how best to partner, and deliver analytics to their customers in those multi-cloud environments. >> Absolutely, so you're obviously speaking with a lot of customers. You're running the solution architect, thr pre-sales, engineering, for your organization. You guys are a partner of Amazon, however your customers are asking you for a multi-cloud approach.
Therefore, I understand in 2020, you guys are exploring a multi-cloud deployments for your customers. You want to talk to me about that? >> Yeah, a little bit, currently we deploy in our own cloud to customers around the world. And, these are customers that are trying to do embedded analytics within a white-label fashion for their products within their cloud environments, and we're actually taking the data, and delivering it to their customers, either B to B, or B to C environment, or internally across business functions.
To have deployed those within a cloud infrastructure, and we host, and deliver that out. But, a lot of customers are now looking for alternatives, and not just running in our cloud, but they want to run a multi-cloud. For example, customers that don't want to run in Amazon, they want to get vendor locked, and so they're looking to also employ Google, or Azure. >> Okay. >> But, they're asking us, looking for you know, it's great that you can host, and manage this for us, but we want to be able to spread that around a little bit, and deploy some of the analytics in other environments.
So, in 2020 part of our roadmap is to be able to deploy in multi-cloud departments. >> Got it. Talk to me a little bit about Sagemaker, I'm excited about Sagemake because there's this growing understanding of mlops in the industry, and there's maybe a push of mlops. I don't want you to get into mlops. What I want to get is your take on Sagemaker, specifically developers not caring about Kubernetes, however how it enables operations to take those workloads to Kubernetes. >> Alright.
So, pretty much before this new invention we had to do double-click on one environment, and then do your modeling in another environment. But, with this IDE, that's the Sagemaker Studio, you're now able to do everything in one place. So, it's like it's one-stop-shop, you know, for building your model, training the model, deploying, and then also monetary. Right in one environment. Which is something I really think is great, and it's going to take production, it's going to reduce the time to production.
>> Got it. So talk to me little bit about, are you working with developers, or are you working with operations? Both. Yeah, so because I work with multiple clients, some clients, some projects, is just with the developers. Some of the projects with the operations. Some projects is both. So, it depends. >> Got it. So, Keith's audience primarily is CTOs, high level, they want to get an understanding. If you want to breakdown machine learning, AI, to a CTO, or a CIO, how would you simplify that?
>> So, we are amplifying what a human can do, and that's teaching a machine to amplify what a human can do. That's in the simplest form. >> Run that back for me one more time. (laughing) >> Teaching a machine to amplify the task that a human being can do. >> Got it. Understood. Kevin, I think you do a good job of explaining Kubernetes. If you wanted to break down Kubernetes to an executive, who may hear the word, and understand that it's necessary, break down for me, simplify it if you don't mind.
Well, Kubernetes and Docker kind of go hand-in-hand. A lot of people will build applications, and they want to deploy them at scale. So, they'll create an application, and they'll containerize it, different pieces of the application. So, different micro-services might be containerized, and deployed. But, to use Kubernetes is a way of scheduling that deployment, and building out clusters, so the application can scale. So you can develop automated pipelines to automatically detect when you've reached a certain threshold on usage, and then automatically scale that cluster.
And, that takes place within Kubernetes. >> Okay. Now, if you just could kind of give us a snapshot as to how GoodData is utilizing Kubernetes. >> Yeah. So GoodData is a cloud (mumbles) application, and it's deployed worldwide. We have data centers all over the world. Our customers ask us to, you know, host and manage their BI applications. That's white label within, they're embedded in their applications. And, so what we do is every week we come up with new patches, and maintenance enhancement release for the product.
And those are rolled out on a weekly basis. And so, we containerize those, and just schedule those to be deployed using Kubernetes. And, then that scales within our ecosystem, so however many instances we have within our ecosystem, and then because the deployment occurs in a service environment, so we manage that through Kubernetes. 8 million users going on there. >> Okay. >> So, we're able to do that uninterrupted, and deploy it worldwide without any interruption, in degradation in SLAs, or performance, because we're leveraging that technology.
>> Got it. Thank you. Thank you. Isi, before we go, do you want to say anything on behalf of Accenture's AWS practice? Anything that you want to communicate to the audience? >> Yes, so the AABG, that's the Accenture AWS Business Group, we are the global partner for AWS. We have over 200 people working full-course, dedicated on this practice, and we deliver best practices across several clients in North America. So, yeah, we're excited to be part of this great move.
>> Nice, nice. Before we go, how can they find you on social media. com. I'm Kevin Young, Roman Stanek's the CTO. And, you know, we're very happy to work with many customers from around the world, and allowing them to derive analytic insights to their customers at scale, and at very high capacity. >> Perfect. Thank you. Isi, before we go, how can they find you on social media. >> So, Isi Idemudia on Twitter, Linked In. You can also reach my management, Chris Wegmann, Chris (mumbles), Accenture Technology, Accenture Cloud.
>> Perfect. Thank you. This is Damani Corbin, sitting in on This is Your CTO Dose. Thank you for listening in. This is AWS re:Invented 2019. Thanks for listening.
