(bright music) >> Welcome back to theCUBE's coverage of AWS re:Invent 2021. I'm John Furrier, your host of theCUBE. We have Michael D´aniello, platform architect at VMware's Carbon Black. Michael, great to see you. We're here at re-Invent virtual hybrid in person. Great to have you on theCUBE. Thanks for coming on. >> Yeah, thanks a lot. Glad to be here. >> So one of the big stories that we're tracking, obviously, is workloads. All cloud for all workloads. Obviously the data is a big part of things, but under the covers and optimizing cloud for the application developers, this modern application movement is more and more at the top of the stack. People just wanting to code. Infrastructure as code. You've seen DevSecOps is a big trend that's driving all new microservices, all new greatness for developers, but still, there's an optimization question. I want to get your thoughts on this, is what you do. Take a minute to explain what your role is at Carbon Black around this cloud optimization. >> Yeah, absolutely. Yeah, so my name is Michael D'aniello. I am a platform architect of VMware Carbon Black. I work across all the different engineering teams. And our main objective is to develop scalable platform tools and that includes, yeah, cloud security, automation pieces, pipelines, cost optimization, like we'll be talking about today, developer enablement tooling and observability tooling. >> One of the big things about instances is that, you know, do I have enough instances? 'Cause honestly, the elastic cloud is amazing, all kinds of new resources there, but talk about the AMD portion of the instances. How do we identify these instances? How to developers understand it, what's in them, and what's the selection criteria? Take us through that whole process of the Amazon web service and the AMD instances. >> Yeah, sure. So essentially, we're leveraging a lot of our instances to run our EKS clusters, which is a managed service for me and for us to run our Kubernetes clusters. And we identify that we can take a bunch of those instances and gain some cost optimization benefits by selecting from Intel to AMD processors. And, you know, initially, we had measured out to be roughly a 10% reduction in cost just for selecting that instance type. But yeah, we actually learned we gained quite a bit more, so. >> John: You know, developers are always like, I want more power, and this is what, you know, the whole idea of Cloud is. Cloud scale has been a big competitive advantage, but also the cost aspect of it. What's the balance between maximizing performance and cost optimization? Because now, you know, people don't want to, you know, they want more power. They also don't want to have a lot of extra spend. And this is kind of one of those things they talk about in Cloud where it's been so successful, cost is important. >> Yeah, yeah, for sure. And it's got to be easy, too, to get that cost optimization benefit. Otherwise, you're spending all your cycles and burning that money there in the human capital and the team and the engineering effort. So luckily, this change is a one-line change. We use Terraform for our automated provisioning, a layer, and we were able to make that one line change and then developers didn't have to make any application changes, which was great. So it was a no-brainer for us to pursue this. >> Talk about the EC2 instances that leverage AMD based process for the EKS, you mentioned that earlier, what is that all about? What's the benefits, what's in it for you guys? >> Yeah, for sure. So essentially, the workloads that are running on these instance types are actual Carbon Black Cloud application. So, all the backend systems that support our customers. And so in that use case, we're, you know, we're spinning up all of our containers that are running our applications and essentially, that's our use case for those instance types. >> How did you come to use the AWS EC2 instances on the AMD? Did you have an evaluation process? Did you just go select it? I mean, take us through that migration aspect of it. >> Yeah, sure, yeah. So originally, we're looking across the board. How can we do better cost optimization, right? And that goes across every different AWS resource, but we targeted this one specifically. We worked alongside with their AWS TAMs and representatives to basically find out, "Hey, is this financially worth the effort?" And we did reach that conclusion with some analysis, basically targeting these instance types and doing some analysis on that cost optimization specifically. And it ended up, you know, being the right thing to target. >> What was the ease of use of the switch? Take us through that. Was it a heavy lift? Was it seamless? Take us through the impact, there, on the move over and what were the results of that? >> Yeah, so I mean, that's the greatest thing. Like I said before, I mean, we had to make just a single line change just to change that instance type in our config and then roll that out across our regions. We did slow roll that in order to make sure that those changes in our development environments didn't make any, you know, performance hits or we didn't run into any snags with the applications themselves. But yeah, I mean, that's the greatest part about the story from my perspective is the ease to migrate over and to switch to these instance types, and then you just immediately gain that cost optimization benefit. >> You know what I love about what your job is, platform architect, that word kind of had a lot of meaning even 10, 15 years ago, but now with the Cloud, it's almost like you're always finagling and managing and massaging and nurturing the infrastructure to enable it. More new things are coming online as well, more high level services. So you've got a fun job and it's always evolving. How do you stay on top of it? What's the impact been for your customers, too, as you start deploying some of these new instance capabilities? Take us through kind of a day in the life of what you do and then what's the impact of customers? >> Yeah, sure. So, you know, like you said, there's quite a bit now to look at. You know, you got to stay on top of different blogs and keep connected with your network to see what your other colleagues are doing across different companies. You know, you can go into conferences like AWS re:Invent, right, to keep on the cutting edge here. But yeah, that's essentially, you know, one of the key aspects is just trying to look at all the different aspects, all the new technologies that are coming out, making sure you're making the right choices there and trying to get the most bang for your buck while you're at it. >> What are some of the big factors that you see in cloud native as you start to look at what customers are doing? Obviously with Kubernetes, you're starting to see that platform develop inside the industry as well as de facto, kind of orchestration layer. But now as customers start to look at it, they want to have more ease of use there, too. At the same time, they don't want to have to do a lot of front end work. They want to get instant benefits in the Cloud, obviously, whether it's from a security standpoint or just rolling out a modern application. Okay, so as having all this infrastructure under the covers, how do you look at that problem and how do you capture that opportunity? >> Yeah, and I think that's why we're seeing a movement here on platform teams. It's kind of a newer terminology, usually a band of developers and SREs come together and say, "Well, we've got a lot of different things to look at. We're onboarding applications to Kubernetes, and we need to make tools so that developers don't have to think much about the transition and the underlying platform." And so that's one of our success metrics on the platform engineering team is just to almost, you know, be non-existent, right? To just have everything flow through our systems and then have just a high ease of use to onboard the applications to the new platform. >> You know, it looks like you have some great success with the AMD based instances. Can I ask you a question? 'Cause I wanted figure this out. How do you identify an AMD based instance when you're making the selections? >> Yeah, sure. It's as easy as just the A after the name. So for us, it was the C5.4XL. And if you want the AMD one, it's just the C5A.4XL. So I guess technically, instead of a one line change, it's actually a one letter change. So, quite easy there. >> Yeah, it's almost like back in the old glory days of command line, one quick update. The customer aspect of this is also important, too. If you don't mind, while I got you here, what are some of the things that you're hearing from your customers, from a performance standpoint, that they're looking for? Obviously, the cost optimization is key, but as they look to deploy more power and more performance, what are some of the things that your customers are looking for from Carbon Black? >> Yeah, so I mean, we are a security company, but we're really a data company because we have, you know, 8,000 customers, we processed over a trillion events per day, we ingress over a hundred terabytes of data per day. And so, our customers need high level performance. And if we can't provide that with low latency, we're not successful. So that's why, you know, performance on the underlying systems that are running our applications is super critical. >> Yeah, you're looking at trailblazer over there. I mean, the work that you guys are doing with the data is amazing. And that's a big theme at re:Invent this year is that data is a huge part. We look at the success of the cloud growth on this, I call gen-two cloud, happening. This whole modern movement is all about how people handle the data at scale, 'cause cloud scales here and now you've got processing all that data, The trailblazing that's going on, there's like this new wave of, I almost called it first-generation trailblazers, but you guys are doing that. What advice would you have for other architects out there and kind of the mainstream enterprises who are like, "Hey, I want to take advantage of the path that you guys have plowed through." What's your advice? >> Yeah, I think one of the key things in a place where we've had a lot of success is creating standards, making sure that we're choosing technology wisely, and making sure that your company isn't building the same solution in silos. And you know, that's a huge pattern that I've seen in my career. And if you can negate that, you're going to be in a great place. So, you know, choose the right technology, container first, cloud native first, push forward, and then make sure that everybody's kind of on that same ship running in the same direction. >> Well, great case study on this AMD based instance migration. Was there any uplift and experience that you've seen on the switch and the performance? Can you just talk about that? What does it mean to upgrade? What benefits are you seeing on the performance you have? >> Yeah, so I didn't hit on this yet and I really wanted to. Yeah, so upfront, the instance itself is 10% cheaper. However, we found out that we had to run far less instances because of that performance increase. So we ended up saving roughly 30% and we've continued to scale out. So at first, it was a couple of hundred instances. Now we're in the thousands and we're going to keep ramping up to over 10 thousands, tens of that. >> John: Let me get this right. So single line change, letter change, instance change. So you get not as many instances, and you save money, so you get cost optimization and higher performance. >> Yep. They say, if it's too good to be true, it's not. But in this case, it actually is. >> So why is it so good in your opinion? What did you discover? What was the big revelation that went down this path? Because that's good value proposition. >> Yeah, for sure. I mean, so initially, we were just chasing that initial BC to 10% and then as we kind of push it forward, we're looking at the metrics, month to month costs and we're actually saying, well, as we kind of swap over from one instance type to another, we're actually paying less. And then once we fully swapped over, it took five or six months to get to the same amount of costs as we continued to scale upward. So it's been a great story. >> It is a great story. It's super nuanced, but it's super important to know these platform benefits. I got to ask you on a personal question, if you don't mind. We love covering Cloud. We've been covering Amazon, it's our ninth year at re:Invent. Just love covering all the action and tech as this just total awesomeness environment. Cloud scale, innovation, capabilities, it's like surfing a big wave. But there's a bigger wave coming and we're seeing it now. I want to get your thoughts on this. As you look to the next big wave, beyond Cloud now, Cloud scale, data, new architecture is rolling out with Edge, basically distributing computing at large scale, and tons of security challenges, right? How do you look at this next big wave coming? Are you staring at it saying, wow, this is going to be huge? And how do you ride that wave? What's your mindset and how do you look at that? >> Well first of all, I'm extremely excited about it. Just the further this thing grows out, there's definitely more complexity, but just a whole slew of fun problems to solve. But when we look at these different problems and solving them at scale across multiple regions, it gets pretty exciting, right? So I can say one example of this is our security of our Cloud, not the security product, and we've developed automation for prevention and auto-remediation in our pipelines. It's been such a success story. And these type of technologies did not exist even a couple of years ago and we've been able to take advantage of them. So, there's going to be a lot more of that where that came from. So, yeah. >> Michael, great work. And again, you're truly a trailblazer, and this is, again, you got to do it. You got to screw your own cloud and stay on the cutting edge and ride that wave. Congratulations on the CostOp cloud optimization and the success with AMD based instances. Congratulations. Thanks. >> Thanks. >> Okay, this is theCUBEs coverage of AWS's re:Invent 2021. I'm John Furrier, your host of theCUBE. Thanks for watching. (inspirational music)

(upbeat music) Welcome back to theCUBE's coverage of AWS re:Invent 2021. I'm John Furrier, host of theCUBE. We have George Watkins, the product marketing manager, cloud gaming and visual cloud at AMD. George, thanks for coming on theCUBE. >> Thank you for having me. >> Love this segment, accelerating game development. AWS cloud, big topic on how the gaming developer environment's changing and how AMD is powering it. Let's get into it. So streaming remote, working remote, flexible collaboration, all powered by the G4ad virtual workstations, it's been a big part of success. Take us through what's going on there. >> Yeah, certainly. So obviously from a remote working perspective, there was a huge impact on collaboration and productivity for many industries out there. But, a collaborative environment like game design, it was even more so. First off, happy to have these big bulky workstation ship to local artists, so they can actually carry on working was a massive nightmare for IT management. Making sure that they have the right hardware, the right resources, the right applications and security. So it was a real mean task. And on top of that, working remotely also brings in other efficiencies when it comes to collaboration. So for example, working on a data sets, as I mentioned before, it's a huge team collaboration effort when it comes to game development, and using the same dataset happens very very often. So if you're actually working remotely and an artist, for example pulled a dataset, from a server, worked on it, then took it back up into the cloud. I'll tell you now, it takes some time to do. And at the same time you might have one or two other artists trying to use that data set. The problem or the big issue that comes here is version control. And essentially because these artists are using the older version, there's creating errors, and keeping that production timed longer. So it's very very inefficient. And then this is where the cloud really comes to end zone. First off the cloud, and then obviously in this case, the AWS cloud, with G4ad instances, really does bring the whole pipeline together. It brings the data sets and the virtual workstations, obviously, as I mentioned, G4ad, as well as all the applications into one place. It's all centralized. And from an IT perspective, this is fantastic. And actually sending out a workstation now is really really simple. It's log in details into an email to your new staff, and there's some really great benefits as well from a staff perspective. Not only are they not tethered to a local workstation, they have the flexibility of work where they need to, and also how they like to. But it's also really interesting about how they work on a day-to-day basis. So a good example of this is, if a artist is using or working on a very very heavy dataset and the configuration from their VM or virtual workstation, isn't up to snuff because of the such a large dataset, all they need to do is call up IT and say, I need more resource. And literally in a couple of minutes time, they can actually have that resource, again, improving that productivity, reducing that time. So it's really really important. And just a final note here as well, with having all that data and all that resource in the cloud, version control tools, really do help bring that efficiency as it's all built into the applications and that data sets really, really exciting staff and ultimately, bring in that productivity and reducing that time and errors down. >> I could see your point too because, when you don't bring it to the cloud, people are going to be bored, waiting for things to happen. And they say I want to take a shortcut. Shortcuts equal mistakes. So, I can see that the G4ad with focus for artists is cool because it's purpose-built for what you're talking about. So take me through how you see the improved efficiencies in the development pipeline with cloud computing around this area because, obviously it makes a lot of sense. Everything's in the cloud, you've got the instances there. Now what happens next? How does the coding all work? What's going on around the game development pipeline? >> 3D applications today, particularly at use in the game industry, I'll be honest, they are still based on legacy hardware. And what I mean by this is that the applications typically require higher CPU Hertz the typically single threaded, maybe some kind of multi threaded functionality there. But generally they are limited by what the traditional workstation has been. And obviously why not? They've been built over the last 10-15 years to access that type of data. Now that is great, but it's not accessing what could be, all the resources that are available in the cloud. And this is what's really really exciting in my part. So ultimately what we're saying is that is that you have this great virtual workstation experience. You have all your applications running on there, you can be efficient, but then there's these really specific and really interesting use cases that aren't accessing the cloud. And I've got a couple of examples, so first off there's a feature inside Unreal 4 engine, called Unreal Swarm. And this feature helps actually reduce the time it takes, in this case and to bake light maps into auto scale, to bake light maps into a game. And this is done by auto scaling, the compiling in AWS cloud. So for example, after making the amends to a light map, we're ready to essentially recompile, but instead of doing this on the local workstation, using the traditional CPU and memory resource, which you would expect to see in a workstation, and actually in this case, it takes around about 50 minutes to do. When you actually use Unreal Swarm, you can, the coordinator as part of this functionality, bursts the requirement or the actual compiling into the cloud. And actually in this case, it's using, like, 10 C5a instances. So these are all CPU high-performance computing instances. And because you have this ability to auto-scale, you actually essentially bring that time, that original 50 minutes, down to 4 minutes. And this type of kind of functionality or this type of task that you would typically see with a 3d artist or with a programmer, basically happens multiple times a day. So when you start factoring in a saving of 45 minutes multiple times a day, it starts really bringing down, the amount of time saved, and obviously the amount of cost saved as well for that artist's time. So it's really really exciting and, certainly something to talk about. >> That's totally cool. I got to ask you since you're here, because it brings up the question that pops into my head, which is okay. What's the state of the art development trends that you're seeing because, on the cloud side, on non gaming world, so shift left to security. You start to see more agile kind of methods around what used to be different modules, right? So you mentioned compiling, acceleration, what's going on in the actual workflows for the developers? What are some of the cool things that you could share that people might not know about that are important? >> Well certainly it's really about finding, those bursty computational expensive and time consuming processes, and actually moving them to the cloud. So really, from a compiling standpoint, they are usually CPU bound. So essentially the GPU does all the work when it comes to the view pole, all that high rendering frames per second, that's what it's really designed for. And it does a very good job with that. But actually the compiling aspect, the compute aspect is all done on the CPU side. And, the work that we've been doing with AWS and the game tech team is actually finding certain ways of actually helping to reduce the compiling nature because ultimately that is always restricted by the amount of calls that you actually have on a local device. So again, another example is there's a company out there called Incredibuild, and they specialize in accelerating the development of that programming code. And obviously in this case, it's the game code. And if an artist, entered a clean source code built on unreal engine full, it would take approximately around about 60 minutes to do on a local machine. However, using the Incredibuild solution to accelerate that type of workload, you can complete it in just 6 minutes. Because again, it's auto scaled out that compiling to several in this case 16 C5a large instances, which essentially reduces all that time for the artist freeing them up to do more stuff. >> And the more creativity is just the classic use case of the clouds, beautiful thing. It's just reminds me of how good this is, because, when you think about what you guys are doing, pushing the envelope for cloud with the creators. gaming is such a state-of-the-art pressure point to make high performance come better. It really is putting a lot of pressure on AMD and everyone else's to get faster and stronger because, it truly is pushing state-of-the-art in general. It's always been that way. If you look at the gaming world. This is a whole 'nother level. I mean, you starting to see that. What's your view on that? If you look at the gaming as a tail sign for the trends and the tech side, better, faster, cheaper processors and speeds and feeds, and how codes work in between processes GPU's and CPU's, all this is cool. All kind of new, if you will. New patterns, new usage, what's your view on that? >> Well certainly, cloud gaming is a really exciting topic and, we believe that cloud gaming with the introduction of various key elements are reading revolutionalize the way that some people are actually using their complaint gamings and interacting with games. And what I mean by this is like, today we can do cloud gaming, it's a fantastic experience. You're usually hardwired, using a broadband connection to actually play those games. And you tend to try and be close to an actual data sensors, to try to reduce that latency. However this is only going to get better with the introduction of 5G coverage and also just, as important edge computing. And because of these two elements, what we're going to be seeing is very high speeds wirelessly, and more importantly, low latency. And this is very important for, that very dynamic cinematic gaming experiences. But not only this, what it can actually do is bring, 4k, 8K gaming to people wirelessly. It can also bring VR and AR experiences wirelessly, and also it can access, these new emerging technologies that are making higher fidelity gaming experiences like hardware retraces. All this can be done with these new technologies. And it's incredibly incredibly exciting. But more importantly, what's really great about this is, from a game publisher perspective, because it's actually helping them simplify their business processes, particularly from a game development standpoint. And actually what I mean by this is, if we take a typical example of what a game developer has to do for a mobile game, there's certain considerations that they need to think about when they actually comes to developing and validate. First off they'll have to understand what type of OS to account for. And actually what type of version of that OS to account for. What type of IPA they're going to be building on. And also finally, what type of resources, are actually on that end point device. So there's a lot of considerations here, and a lot of testing. So ultimately a lot of work to get that game out, to those gamers who might be on a couple of these different mobile platforms. However, when it comes to game streaming, it really does kind of change all this because ultimately what the game developer is actually doing is that they're developing and they're validating on one source. And that is going to be the server that is essentially pairing that game streaming service. Because how game streaming works is that we essentially trans code the actual game via H.264 to a software client on any end point device. So this could be those mobile devices I just mentioned. It can also be TVs, it could be consoles, it can be even low powered laptops. And what's very exciting is that, from an end user perspective, they're getting the ultimate in gaming experiences and usually these types of solutions are traditionally subscription-based. So you're actually reducing the requirement of this kind of high-end thousands of dollars gaming solution or simply a high-end next gen console. All of this is actually been given to you and delivered as part of a game streaming service. So it's very very exciting and, certainly we can see the adoption on both the game development side, as well as the gamer's side. That's a great way to put an end to this awesome segment. I think that business model innovation around making it easier, and making it better to develop environment, that's just how they work. So that's good, check. But really the business model here, the gaming as a service, you're making it possible for the developer and the artist to see an outcome faster. That's the cloud way. >> Thank you >> And they doubled down on success and they could do that. So again, this is all new and exciting and certainly the edge and having data being processed at the edge as well. Again, all this is coming in to create more good choice. Thank you so much for coming on and sharing that insight with us from the AMD perspective. And again, more power, more speed, we always say, no one's going to complain, they get more compute, that's what I always. >> Absolutely absolutely. >> Thanks for coming I appreciate it. >> Thank you. >> theCUBE coverage here at AWS re:Invent 2021. I'm John Furrier host of theCUBE. Thanks for watching. (upbeat music)

(bright music) >> Welcome back to theCUBE's coverage of AWS re:Invent 2021. I'm John Furrier, your host of theCUBE. We have Michael D´aniello, platform architect at VMware's Carbon Black. Michael, great to see you. We're here at re-Invent virtual hybrid in person. Great to have you on theCUBE. Thanks for coming on. >> Yeah, thanks a lot. Glad to be here. >> So one of the big stories that we're tracking, obviously, is workloads. All cloud for all workloads. Obviously the data is a big part of things, but under the covers and optimizing cloud for the application developers, this modern application movement is more and more at the top of the stack. People just wanting to code. Infrastructure as code. You've seen DevSecOps is a big trend that's driving all new microservices, all new greatness for developers, but still, there's an optimization question. I want to get your thoughts on this, is what you do. Take a minute to explain what your role is at Carbon Black around this cloud optimization. >> Yeah, absolutely. Yeah, so my name is Michael D'aniello. I am a platform architect of VMware Carbon Black. I work across all the different engineering teams. And our main objective is to develop scalable platform tools and that includes, yeah, cloud security, automation pieces, pipelines, cost optimization, like we'll be talking about today, developer enablement tooling and observability tooling. >> One of the big things about instances is that, you know, do I have enough instances? 'Cause honestly, the elastic cloud is amazing, all kinds of new resources there, but talk about the AMD portion of the instances. How do we identify these instances? How to developers understand it, what's in them, and what's the selection criteria? Take us through that whole process of the Amazon web service and the AMD instances. >> Yeah, sure. So essentially, we're leveraging a lot of our instances to run our EKS clusters, which is a managed service for me and for us to run our Kubernetes clusters. And we identify that we can take a bunch of those instances and gain some cost optimization benefits by selecting from Intel to AMD processors. And, you know, initially, we had measured out to be roughly a 10% reduction in cost just for selecting that instance type. But yeah, we actually learned we gained quite a bit more, so. >> John: You know, developers are always like, I want more power, and this is what, you know, the whole idea of Cloud is. Cloud scale has been a big competitive advantage, but also the cost aspect of it. What's the balance between maximizing performance and cost optimization? Because now, you know, people don't want to, you know, they want more power. They also don't want to have a lot of extra spend. And this is kind of one of those things they talk about in Cloud where it's been so successful, cost is important. >> Yeah, yeah, for sure. And it's got to be easy, too, to get that cost optimization benefit. Otherwise, you're spending all your cycles and burning that money there in the human capital and the team and the engineering effort. So luckily, this change is a one-line change. We use Terraform for our automated provisioning, a layer, and we were able to make that one line change and then developers didn't have to make any application changes, which was great. So it was a no-brainer for us to pursue this. >> Talk about the EC2 instances that leverage AMD based process for the EKS, you mentioned that earlier, what is that all about? What's the benefits, what's in it for you guys? >> Yeah, for sure. So essentially, the workloads that are running on these instance types are actual Carbon Black Cloud application. So, all the backend systems that support our customers. And so in that use case, we're, you know, we're spinning up all of our containers that are running our applications and essentially, that's our use case for those instance types. >> How did you come to use the AWS EC2 instances on the AMD? Did you have an evaluation process? Did you just go select it? I mean, take us through that migration aspect of it. >> Yeah, sure, yeah. So originally, we're looking across the board. How can we do better cost optimization, right? And that goes across every different AWS resource, but we targeted this one specifically. We worked alongside with their AWS TAMs and representatives to basically find out, "Hey, is this financially worth the effort?" And we did reach that conclusion with some analysis, basically targeting these instance types and doing some analysis on that cost optimization specifically. And it ended up, you know, being the right thing to target. >> What was the ease of use of the switch? Take us through that. Was it a heavy lift? Was it seamless? Take us through the impact, there, on the move over and what were the results of that? >> Yeah, so I mean, that's the greatest thing. Like I said before, I mean, we had to make just a single line change just to change that instance type in our config and then roll that out across our regions. We did slow roll that in order to make sure that those changes in our development environments didn't make any, you know, performance hits or we didn't run into any snags with the applications themselves. But yeah, I mean, that's the greatest part about the story from my perspective is the ease to migrate over and to switch to these instance types, and then you just immediately gain that cost optimization benefit. >> You know what I love about what your job is, platform architect, that word kind of had a lot of meaning even 10, 15 years ago, but now with the Cloud, it's almost like you're always finagling and managing and massaging and nurturing the infrastructure to enable it. More new things are coming online as well, more high level services. So you've got a fun job and it's always evolving. How do you stay on top of it? What's the impact been for your customers, too, as you start deploying some of these new instance capabilities? Take us through kind of a day in the life of what you do and then what's the impact of customers? >> Yeah, sure. So, you know, like you said, there's quite a bit now to look at. You know, you got to stay on top of different blogs and keep connected with your network to see what your other colleagues are doing across different companies. You know, you can go into conferences like AWS re:Invent, right, to keep on the cutting edge here. But yeah, that's essentially, you know, one of the key aspects is just trying to look at all the different aspects, all the new technologies that are coming out, making sure you're making the right choices there and trying to get the most bang for your buck while you're at it. >> What are some of the big factors that you see in cloud native as you start to look at what customers are doing? Obviously with Kubernetes, your starting to see that platform develop inside the industry as well as defacto, kind of orchestration layer. But now as customers start to look at it, they want to have more ease of use there, too. At the same time, they don't want to have to do a lot of front end work. They want to get instant benefits in the Cloud, obviously, whether it's from a security standpoint or just rolling out a modern application. Okay, so as having all this infrastructure under the covers, how do you look at that problem and how do you capture that opportunity? >> Yeah, and I think that's why we're seeing a movement here on platform teams. It's kind of a newer terminology, usually a band of developers and SREs come together and say, "Well, we've got a lot of different things to look at. We're onboarding applications to Kubernetes, and we need to make tools so that developers don't have to think much about the transition and the underlying platform." And so that's one of our success metrics on the platform engineering team is just to almost, you know, be non-existent, right? To just have everything flow through our systems and then have just a high ease of use to onboard the applications to the new platform. >> You know, it looks like you have some great success with the AMD based instances. Can I ask you a question? 'Cause I wanted figure this out. How do you identify an AMD based instance when you're making the selections? >> Yeah, sure. It's as easy as just the A after the name. So for us, it was the C5.4XL. And if you want the AMD one, it's just the C5A.4XL. So I guess technically, instead of a one line change, it's actually a one letter change. So, quite easy there. >> Yeah, it's almost like back in the old glory days of command line, one quick update. The customer aspect of this is also important, too. If you don't mind, while I got you here, what are some of the things that you're hearing from your customers, from a performance standpoint, that they're looking for? Obviously, the cost optimization is key, but as they look to deploy more power and more performance, what are some of the things that your customers are looking for from Carbon Black? >> Yeah, so I mean, we are a security company, but we're really a data company because we have, you know, 8,000 customers, we processed over a trillion events per day, we ingress over a hundred terabytes of data per day. And so, our customers need high level performance. And if we can't provide that with low latency, we're not successful. So that's why, you know, performance on the underlying systems that are running our applications is super critical. >> Yeah, you're looking at trailblazer over there. I mean, the work that you guys are doing with the data is amazing. And that's a big theme at re:Invent this year is that data is a huge part. We look at the success of the cloud growth on this, I call gen-two cloud, happening. This whole modern movement is all about how people handle the data at scale, 'cause cloud scales here and now you've got processing all that data, The trailblazing that's going on, there's like this new wave of, I almost called it first-generation trailblazers, but you guys are doing that. What advice would you have for other architects out there and kind of the mainstream enterprises who are like, "Hey, I want to take advantage of the path that you guys have plowed through." What's your advice? >> Yeah, I think one of the key things in a place where we've had a lot of success is creating standards, making sure that we're choosing technology wisely, and making sure that your company isn't building the same solution in silos. And you know, that's a huge pattern that I've seen in my career. And if you can negate that, you're going to be in a great place. So, you know, choose the right technology, container first, cloud native first, push forward, and then make sure that everybody's kind of on that same ship running in the same direction. >> Well, great case study on this AMD based instance migration. Was there any uplift and experience that you've seen on the switch and the performance? Can you just talk about that? What does it mean to upgrade? What benefits are you seeing on the performance you have? >> Yeah, so I didn't hit on this yet and I really wanted to. Yeah, so upfront, the instance itself is 10% cheaper. However, we found out that we had to run far less instances because of that performance increase. So we ended up saving roughly 30% and we've continued to scale out. So at first, it was a couple of hundred instances. Now we're in the thousands and we're going to keep ramping up to over 10 thousands, tens of that. >> John: Let me get this right. So single line change, letter change, instance change. So you get not as many instances, and you save money, so you get cost optimization and higher performance. >> Yep. They say, if it's too good to be true, it's not. But in this case, it actually is. >> So why is it so good in your opinion? What did you discover? What was the big revelation that went down this path? Because that's good value proposition. >> Yeah, for sure. I mean, so initially, we were just chasing that initial BC to 10% and then as we kind of push it forward, we're looking at the metrics, month to month costs and we're actually saying, well, as we kind of swap over from one instance type to another, we're actually paying less. And then once we fully swapped over, it took five or six months to get to the same amount of costs as we continued to scale upward. So it's been a great story. >> It is a great story. It's super nuanced, but it's super important to know these platform benefits. I got to ask you on a personal question, if you don't mind. We love covering Cloud. We've been covering Amazon, it's our ninth year at re:Invent. Just love covering all the action and tech as this just total awesomeness environment. Cloud scale, innovation, capabilities, it's like surfing a big wave. But there's a bigger wave coming and we're seeing it now. I want to get your thoughts on this. As you look to the next big wave, beyond Cloud now, Cloud scale, data, new architecture is rolling out with Edge, basically distributing computing at large scale, and tons of security challenges, right? How do you look at this next big wave coming? Are you staring at it saying, wow, this is going to be huge? And how do you ride that wave? What's your mindset and how do you look at that? >> Well first of all, I'm extremely excited about it. Just the further this thing grows out, there's definitely more complexity, but just a whole slew of fun problems to solve. But when we look at these different problems and solving them at scale across multiple regions, it gets pretty exciting, right? So I can say one example of this is our security of our Cloud, not the security product, and we've developed automation for prevention and auto-remediation in our pipelines. It's been such a success story. And these type of technologies did not exist even a couple of years ago and we've been able to take advantage of them. So, there's going to be a lot more of that where that came from. So, yeah. >> Michael, great work. And again, you're truly a trailblazer, and this is, again, you got to do it. You got to screw your own cloud and stay on the cutting edge and ride that wave. Congratulations on the CostOp cloud optimization and the success with AMD based instances. Congratulations. Thanks. >> Thanks. >> Okay, this is theCUBEs coverage of AWS's re:Invent 2021. I'm John Furrier, your host of theCUBE. Thanks for watching. (inspirational music)