PurePerformance - Optimizing Your Migration to AWS with Carmen Puccio AWS

Episode Date: January 30, 2019

Carmen Puccio, Principal Solutions Architect at AWS, talks to us about the Dynatrace Managed AWS quick-start program we created with him, modeling your monolith as a microservices platform, some of the Dynatrace announcements from Perform as well as Carmen’s snow adventure with guest of the show Mandus Momberg

Coming to you from Dynatrace Perform in Las Vegas, it's Pure Performance! Hey! Welcome back. Hey Mark, we're back here at Dynatrace, and we have Carmen Puccio. Puccio. Carmen Puccio from AWS. Yes. Welcome to the podcast. Thank you for having me.

This is great. How's your Dynatrace experience so far? It's been good. It's been good. I've been hanging out in the conference center here, and I'm doing booth duty today. And one of my colleagues, he actually just gave a session. So he said it was super interesting. He had a lot of folks in there.

So a little upset I couldn't speak this week, but here I am. I'm talking to you guys. So you're speaking. Yeah, yeah. That's very cool. And when did you get in here? Did you get in here yesterday? I flew in Sunday, actually.

I was in Seattle. So I'm based out of New York, and I was in Seattle all last week doing a little AWS internal planning for this year. And I managed to squeeze a little snowboarding over the weekend out in Washington state, and here I am. So awesome. That sounds cool. Yeah, yeah. Yeah, it sounds cool. So what do you got going on with, just before we started this,

you said you had a couple of projects you were working on with Dynatrace and AWS? Yeah, yeah, absolutely. So I mean, to set a little context, I've been with AWS for almost three years now. Dynatrace has been one of my partners since day one. I started out as a partner solutions architect with AWS, and I curated and worked with our mass migration partners.

So I did the migration competency inside AWS, which Dynatrace is actually a member of. And I've met the team over the years. So whether it be Franz or Andreas or Andy, great guys to work with. And we've kind of inserted Dynatrace through various aspects of like a migration journey. Right, right.

How you can use them as part of your strategic planning as you build your portfolio. What do I have, when do I move it? And then at the same time, the inverse side of that, how do I use Dynatrace to kind of break down my application once I move it, make sure it's performing well, or maybe I want to optimize it, right? So like there was all kinds of opportunities

over the better part of two and a half years to work with Dynatrace in that migration scenario, right? And it's been cool. In the last year and a half, give or take, I've done two things recently with Dynatrace. We've built a quick start. So for customers that want to manage Dynatrace themselves, as opposed to your SaaS offering,

they want to deploy it into an AWS environment, we have something called the AWS Quickstart program, which is essentially a templatized version of a production ready application. A lot of our ISV partners have taken advantage of it because customers ask them, how can I have a prescriptive, production ready, infrastructure as code, you know,

way to drop your software into my AWS account. And Dynatrace, myself and Chris out of Austria, we built Quickstart over about a year, and it's been popular from what I understand. So we did that, and then Andy and I were doing the Break the Monolith story, which I'm sure you guys have seen. Oh yeah, yeah, part of the self-service, I'm trying to, there's the unbreakable pipeline piece.

That was one, that was one, yeah. So Break the Monolith is where you can, you can take your monolithic application, and you can create virtual microservices by creating custom services and see how it's going to flow so that you know whether or not is this going to be a good idea to create a microservice out of this or break this apart and all,

and you really get to model it. And then I think Sonya, they had a hot day session that was using, I think the self-service or the unbreakable pipeline had the Lambda integration I think Andy was working on. Okay, okay. So Andy did an unbreakable pipeline had the Lambda integration I think Andy was working on. Oh yeah, so Andy did an unbreakable pipeline, a CICD pipeline

with the AWS what's the pipeline? Code pipeline. Code pipeline, there you go. It's a code pipeline, I don't know if you know what it's called. What's your code pipeline called? But he would then take data from Dynatrace, feed it into a Lambda function and then do

self-healing. So a lot of really, really cool integrations going on in there. Yeah, yeah. So how do people get into, you were talking about the migration before. How do people get into this whole migration piece of it? Yeah. Yeah, so I mean, a lot of times, too, if you think about it,

a lot of customers want to move their applications over. But at the same time, maybe they have a relationship with Dynatrace already. Right? So you guys have your SaaS offering. And a lot of times, customers want to run it themselves, right? They're on top of EC2 or whatever it may be, and that might be part of their migration. They might want to actually pick up their Dynatrace implementation and figure out how to run it on AWS themselves.

It was about a year-ish ago, I met with the Dynatrace team at RSKO, and we came up with the concept of building a quick start, right? So the AWS quick start program is essentially a way for you to deploy production ready applications on top of AWS. So think about like, you know, you're an ISV partner and people ask you all the time, how do I get started? How do I drop this in? And at the same time, you know, I don't want to build this by hand. I just want to run a CloudFormation script or whatever it may be. AWS works with a partner to build out this production ready implementation. So I worked with Chris from the Dynatrace team

and we built a quick start and it was launched, I want to say it was launched I think somewhere back in September, shortly before our re-invent. But yeah, it was a great journey and I'm glad there's some traction on it for you guys. Yeah, that's excellent. And in terms of moving people over

and helping customers just from the AWS side, bring the infrastructure over, you all have an organization. And I happen to know this because a few weeks ago we had an interview with Mandus Lomborg. My colleague, yes. Wonderful guy. Wonderful guy. Good skier. He skied with me last weekend.

Is he a snowboarder or a skier? No, I'm the snowboarder, he's the skier. There you go. He was very impressive. He was very impressive. First time. His first time? For me to for his first time for me to go with him with him yeah he took me up to Washington to a place called Crystal Mountain really okay yeah yeah so so he was talking about the six stars and I think he made up a seven R

and I know but I hadn't heard about the seven hours before that yeah anything is I was talking to another customer who went through that program yep Samba safety mm-hmm and they went through that whole program and moved everything. And the guy, we were talking about all these different pieces of it, and we're like, oh, that's really interesting. And then next week we interview Mandus, and he's talking about these six R's. I'm like, wait, if I would have known, I could have put this in much better context. He's getting all the credit here because I'm going to say he coined it, right? So the six R's is something that we publicize. It's in all of

our public-facing material. We're approaching a new R internally, really. This is something that we're starting to breach out and we're starting to teach our consulting partners. And essentially it's like, how do I pick my application up and move it into a container as part of my migrations? Say that three times fast. But if you think about it, we pick these applications up, they're lift and shift. We have the server migration service. We have great tools like CloudEndure, which do this lift and shift model. But at the same time, too, is that always necessary, right? The nature of a container, the ability

to build it fast and fail fast, learn from your mistakes, it could be used as part of your migration strategy. Simple web server, maybe just serving NGINX or IIS or whatever it may be. Maybe you don't want to go down that traditional EC2 route. Maybe you want to put that inside of a container as part of your move. Smaller footprint, you know, easier to iterate on. And again, you fail, oh well, tear it down, start all over again, you know, check it in,

and next thing you know, you have a new version of it running. So we're approaching that as our seventh R. It's not official, but it's something that we're starting to explore. We're starting to work with our consulting partners with. And at the same time, too, you know, we've chatted with Dynatrace quite a bit here. How can Dynatrace help show those metrics about, all right, I had the application before. I've stuck it into a container. I've deployed it.

Now I want to do a comparison model. I want a before and after, right? Because, again, just because it's in a container doesn't mean that you still have to monitor it. You still have to measure it because you want it to be optimized and running at least on par, if not better, than what it was on-prem. Right. Excellent. Did you get to see any of the keynote this morning?

No, no. I've been in the booth all morning. So anything interesting come out of it? Yeah, there were two of the bigger announcements. One was around a developer edition. So we have a developer for life edition where they're going to have, I guess it's going to be hosted by us.

They're going to have access to, I don't know how, stripped down. It might just be one agent. I don't know what the restriction will be. Very interesting. But there's an entire community built around it with videos help. They're going to try to get all the developers to be communicating with each other and sharing techniques.

So besides learning how to maybe monitor and look at performance, I think it's also a playground or a sandbox for them to try some of these other integrations. That's a really big one, too. I think we call that the shift left mentality, where we're putting a little bit more responsibilities about how to operationalize an application in the developer's hand.

Containers and serverless really push that envelope. And the fact that you guys are going with that strategy and put an offering out, I think that's fantastic. Yeah, no, it's going to be free too, which is crazy. That's even better. So they'll get to learn the ropes. They'll get to figure everything out that they need to. So once they go to really implement,

they have it all worked out and they can build it right into their code and check it in then, right? Yeah. And then the other piece was the Open AI. AI, not API. I don't want to be like Brandy. But it's the AIOps.

Yes, AIOps. And also the other three-letter acronym is the Autonomous Cloud Management. Right, Autonomous Cloud Management. So ACM. ACM, okay. But the big piece of this now is that the AI, we've always been able to ingest data via API into Dynatrace. Now the AI is going to look at that data as well

as part of the causation engine. So if you have a load balancer or any other appliance that's feeding data into it, that's going to now be, that we can't put an agent on, right? That's going to now be considered as whether or not that's a root cause. So it's going to really open the door for that. MARK MANDELMANN, How does AI work that way internally to AWS?

You guys would be able to share that data? So that's something I'm not necessarily well versed on, so I can't go super deep there. But yeah, I mean, obviously there's opportunities to talk with other SA's that specialize in the AI space. We have a whole wing of SA's that work with partners such as Dynatrace.

And you probably already have one that I've just never met. Yeah, yeah. And I imagine we're going to be able to now, because also in the past we had the CloudWatch metrics. So I'm imagining now we're going to be able to do more with those metrics than just display them, actually get results based on that, which is really exciting. The root cause analysis, sorry, go ahead.

I think the natural integration point is really at machine data, at logs. And certainly AWS makes it really easy to send that machine data to an aggregator someplace else. And that aggregator can certainly be Dynatrace and then the AI engine. Yeah, I mean, we talk about the context of centralized log management as part of our migration strategy. So I think one of the common questions we get

is like, how do I start? I'm new to the cloud. I don't know necessarily where to begin. So besides the part that I talked about, get an understanding of what the application's doing. That's a very key piece of your migration strategy. But at the same time too, you have to think about the operational constructs that it takes

to actually operate at scale. So a lot of times I tell customers to think about things like centralized blog management. Think about things, how am I going to do a multi-account structure inside of AWS? You know, we recently just did the control tower release or a preview announcement where essentially it's allowing customers to have almost like a vending machine style account creation for them. So you know it's not

necessarily trying to crunch everything into a singular account, you want to break things out where you have a shared services account, maybe you have an account specifically for security and application accounts, so on and so forth. But the key theme there is you want to think about log aggregation and how do I have a shared services account, maybe you have an account specifically for security, an application account, so on and so forth. But the key theme there is you want to think about log aggregation and how do I actually take all of those logs, put them to a centralized place so your operations teams and hopefully your developers

can go in and check out and see what's going on at any given time, right? Yep, absolutely. One of the key questions I will have is, especially with the developer edition, I do always want to ask, you find these developer free edition for life announcement, but also how do people get started? Yeah.

Like what do you recommend to people who are still just learning about the cloud? Yeah. I mean, do they just start at running? Yeah, the credit card scenario, right? The developer spins up off his credit card, next thing you know it's the production app. No, I mean along the similar theme

that I was just talking about, all those operational constructs, but the one thing that I always talk about, and we have a whole program around this. We have what's called a migration acceleration program at AWS, where either AWS professional services or consulting partners will go in and work with a customer to help them migrate workloads and teach them how to do it. And almost always an ISV partner such as yourselves or one maybe of our other migration competency partners

are in there helping do a certain function. Maybe it's the monitoring and logging piece, whatever it may be. Right. But the thing that we always tell them is, is don't try and move your biggest workload first, right? We make a lot of tools that make it very easy for you to do things like, again,

pick up that application and move it. Maybe it's in the context of our server migration service or a database migration service or whatever it may be. But think about more, how do I operate, right? And take a smaller application, it's indicative of your workloads and maybe it has a similar deployment strategy or whatever it may be,

that actually has active development cycles and try and move that one first after you've thought about those operational constructs. Again, how am I doing centralized log management? What is my security model look like? How am I going to have users federate in? Stuff like that. Then try and move that application because in a way, it's almost like a testbed about all of your other prior strategy aspects. If you think about everything else that you put into the migration in those key areas,

in those operational constructs, that first application you deploy essentially validates all that. If that works, then in theory, you're supposed to be able to operate at scale and that's how we preach our migration acceleration program that's cool yeah very very cool yeah anything else you're looking forward to doing while you're out here uh-huh no no gambling no gambling no I don't have the best show when it comes that I actually was here with with my girlfriend during the reinvent time and I saw a show over at the MGM it was one of the lists Cirque du Soleil's was that the Beatles?

No, no. They were like in some sort of like, it was almost like pirate themed in the southeast. It was amazing. It was amazing. They were running around in the crowd and everything. So no, I'm going to go home on Thursday. So no shows this week. But yeah, last time I was here, I saw that and anybody who's

never seen it, check it out. It was pretty cool. Yeah, and you went to see the Beatles show, right? Yeah, I've seen several of them. Yeah, well, I mean, for me, the best thing about the Beatles show was they have this phenomenal sound system in there and you're just hearing the Beatles music remixed because I don't know if you've ever heard the soundtrack there. It's like a mega mix almost. Yeah.

Through this huge, huge sound system, really loud. It's an exciting experience for that. But the show was good. It was a little bit more dancing than acrobatics at all. And that's the one at the Mirage, right? Yeah. Yeah, yeah. I just want to point out

Donnie and Marie are at the Flamingo. Oh. Well, boy, that takes me back. Yeah. A young kid. And on that note. Yes, on that note. Awesome.

Yeah, thank you guys for having me. Yeah, it's been a great conference so far. And again, the partnership with you guys has been fantastic. So more to come. Yeah, thank you. Thanks for being here. Cheers.