 From the SiliconANGLE Media Office in Boston, Massachusetts, it's theCUBE. Now, here's your host, Stu Miniman. Hi, I'm Stu Miniman and welcome to the first interview of theCUBE in our Boston area of studio for 2020. And to help me kick it off, Jeremy Daley, who is the host of serverless chats, as well as runs the serverless day, Boston. Jeremy saw you at Reinvent way back in 2019. And we'd actually had some of the people in the community that were like, hey, I think you guys like actually live and work right near each other. And you're only about 20 minutes away from our office here. So thanks so much for making the long journey here and not having to get on a plane to join us here. Well, thank you for having me. All right, so as Calvin from Calvin and Hobb says, it's a new decade, but we don't have any base on the moon. We don't have flying cars that general people can use. But we do have serverless and robot vacuum cleaners. We do have robot vacuum cleaners. We're run by serverless, that's a matter of fact, yes. A CUBE alum on the program would be happy that we do get to mention there. So yeah, serverless, there are things like the iRobot as well as Alexa are some of the things that people, it's usually when I'm explaining to people what this is and they don't understand it, it's like, oh, you've used Alexa. Well, those are the functions underneath. And you think about how these things turn on and off a little bit like that. But maybe we don't need to get into long ontological discussions or everything, but you're a serverless hero. So give us a little bit what you're hearing from people. What are some of the exciting use cases out there and where is serverless being used in that maturity today? Yeah, I mean, well, so the funny thing about serverless and the term serverless itself, and I do not want to get into a long discussion about this obviously, I actually wrote a post last year that was called stop calling everything serverless because basically people are calling everything serverless. So really what I look at it as is something where it just makes it really easy for developers to abstract away that backend infrastructure and not having to worry about setting up Kubernetes or going through the process of setting up virtual machines and installing software is just, a lot of that stuff is kind of handled for you. And I think that has enabled a lot of companies, especially startups is a huge market for a serverless, but also enterprises enabled them to give more power to their developers and be able to look at new products that they want to build, new services they want to tackle, or even old services that they need to, that may have some stability issues or things like long running ETL tasks and other things like that that they've found a way to sort of find the peripheral edges of these monolithic applications or these mainframes that they're using and find ways to run very small jobs using functions as a service, something like that. And so I see a lot of that, I think that is a big use case. You see a lot of large companies doing. Obviously people are building full-fledged applications. So yes, the web-facing user application, certainly a thing. People are building APIs, you've got API Gateway, they just released the new HTTP API, which makes it even faster. To run those sort of things, this idea of cold starts in AWS trying to get rid of all of that stuff with the new VPC networking and some of the other things that they're doing there. So you have a lot of those types of applications that people are building as well, but it really runs the gamut. I mean, there are things all across the board that you can do and pretty much anything you can do with the traditional computing environment, you can do with a serverless computing environment. And obviously that's focusing quite a bit on the functions as a service side of things, which is a very tiny part of serverless if you wanna look at it, sort of the broader picture of the serviceful or managed services type approach. And so that's another thing that you see where you used to have companies setting up MySQL databases and clusters trying to run these things, or even worse, Cassandra rings, right? Trying to do these things and manage this massive amount of infrastructure just so that they could write a few records to a database and read them back for their application. And that would take months sometimes for them to get it set up and even more time to try to keep running them. And so this sort of revolution of managed services and all these things that we get now, whether that be things like manage Elastase search or Elastase search cloud doing that stuff for you, big table and DynamoDB and managed Cassandra, whatever those things are, are just making it a lot easier for developers to just say, hey, I need a database and okay, here it is. And I don't have to worry about the infrastructure at all. So I think you see a lot of people and a lot of companies that are utilizing all of these different services now and essentially are no longer trying to reinvent the wheel. So a couple of years ago, I was talking to Andy Jassy on an interview at theCUBE and he said, if I was to build AWS today, I would have built it on serverless. And what I've seen over the last two or three years or so, Amazon is rebuilding a lot of their servers underneath. It's very interesting to watch that platform changing. I think it's had some ripple effect dynamics inside the company because Amazon's very well known for their two pizza teams and therefore all of their products are there. But I think it was actually in a conversation with you we're talking about in some ways, this new way of building things is a connecting fabric between the various groups inside of Amazon. So I love your viewpoint that we shouldn't just call everything serverless, but in many ways, this is a revolution and a new way of thinking about building things. And therefore, there are some organizational and dynamical changes that happen for an Amazon but for other people that start using it. Yeah, well, I mean, I actually was having a conversation with a Jay Nair who's one of the product owners for Lambda. And he was saying to me, he's like, well, how do we sell serverless? Like how do we tell people, this is the next way to do things? I said, just it's the way, right? And then Amazon has realized this and part of the great thing about dog fooding your own product is that you say, okay, I don't like the taste of this bit, so we're gonna change it to make it work. And that's what Amazon has continued to do. So they run into limitations with serverless just like us early adopters run into limitations and they say, well, how do we make it better? How do we fix it? And they've always been really great to it listening to customers. I complain all the time. There's other people that complain all the time that say, hey, I can't do this. And they say, well, what if we did it this way? And out of that, you get things like Lambda destinations and all different types of ways. You get a vent bridge. You get different ways that you can solve those problems and that comes out of them using their own services. So I think that's a huge piece of it, but that helps enable other teams to get past those barriers as well. Jeremy, I'm gonna be really disappointed if in 2020, I don't see a T-shirt from one of the serverless days with a Mandalorian on it saying serverless, this is the way. So great, great, great marketing opportunity. And I do love that because like some of the other spaces, we're not talking about a point product or a simple thing we do. It is more the way of doing things. It's just like, I think about cybersecurity. Yes, there are lots of products involved here, but this is more of, it's a methodology and it needs to be fully thought of across the board as to how you do things. So let's take it a little bit. At Reinvent, when I went to the serverless gathering, it was serverless for everyone? Serverless for everyone, yes. And it was, hey, serverless isn't getting talked, serverless isn't as front and center as some people might think. Some people on the outside look at this and they say, oh, serverless, those people, they have religion and they go so deep on this. But I thought Tim Wagner had a really good blog post that came out right after Reinvent. And what we saw is not only is Amazon changing underneath the way things are done, but it feels that there's this bridging between what's happening in Kubernetes, you see where Fargate is and Firecracker, and serverless, and help us squint through that and understand a little bit what you're seeing, what your take was at Reinvent, what you like, what you're hoping to see, and how does that whole containerization and Kubernetes wave intersect with what we're doing in serverless? Yeah, well, I mean, for some reason people like Kubernetes. And I honestly, I don't think there's anything wrong with it, I think it's a great container orchestration system, I think containers are still a very important part of the workloads that we're putting into a cloud. I don't know if I would call them cloud native exactly, but I think what we're seeing, or at least what I'm seeing that I think Amazon is seeing, is they're saying people are embracing Kubernetes and they're embracing containers. And whether or not containers are ephemeral or long running, which I read a statistic at some point that was something like 63% of containers, so even running on Kubernetes or whatever, run for less than 10 minutes, right? So basically most computing that's happening now is fairly ephemeral, and as you go up, I think it's 15 minutes or something like that, it's like 70% or 90% or whatever that number is, I totally got that wrong. But I think what Amazon is doing is they're trying to basically say, look, we were trying to sell serverless to everyone. We were trying to sell this idea of, look, manage services, manage compute, the idea that we can run even containers as close to the metal as possible with something like Fargate, which is what Firecracker is all about, being able to run virtual machines, basically almost right on the metal, right? I mean, it's so close that there's no levels of abstraction that get in the way and slow things down. And even though we're talking about milliseconds or microseconds, it's still something and there's efficiencies there. But I think what they looked at is they said, look, we are not Apple, we can't kill Flash just because we say we're not gonna support it anymore. And I think you actually mentioned this to me in the past where the majority of Kubernetes clusters that were running in the public cloud were running in Amazon anyways. And so you had using virtual machines, which are a great technology, but are 15 years old at this point. Even containerization, there's more problems to solve there, but getting to the point where we say, look, you wanna take this container, this little bit of code or this small service and you wanna just run this somewhere. Why are we spinning up virtual containers? Why are we using 15 or 10 year old technology to do that? And Amazon's just getting smarter about it. So Amazon says, hey, if we can run a Lambda function on Firecracker and we can run a Fargate container on Firecracker, why can't we run, can't create some pods and run some pods for Kubernetes on it? They can do that. And so I think for me, I was disappointed with keynotes because I don't think there was enough serverless talk, but I think what they're trying to do is they're trying to, and this is if I put my analyst hat on for a minute, I think they're trying to say, the world is at Kubernetes right now and we need to embrace that in a way that says we can run your Kubernetes for you a lot more efficiently and without you having to worry about it than if you use Google or if you use some other cloud provider or if you run on-prem, which is, I think their biggest competitor to Amazon is still on-prem, especially in the enterprise world. So I see them as saying, look, we're gonna focus on Kubernetes but as a way that we can run it our way. And I think that's why Fargate and Kubernetes or the Kubernetes for Fargate or whatever that new product is, it's too many product names at AWS. But I think that's what they're trying to do and I think that was the point of this is to say, listen, you can run your Kubernetes and Claire Ligori, who showed that piece at the keynote, that Werner's keynote that was basically how quickly Fargate can scale up the individual containers of Kubernetes as opposed to launching new VMs or EC2 instances. So I thought that was really interesting but that's my overall take is just that they're embracing that because I think that's where the market is right now and they just haven't yet been able to sell this idea of serverless, even though you're probably using it with a bunch of things anyways, at least what they would consider serverless. Yeah, to depart a little bit from the serverless for a second, talk about multi-cloud. It was one of the biggest discussions we had in 2019 when I talked to customers that are using Kubernetes, one of the reasons they tell me they're doing it is, well, I love Amazon, I really like what I'm doing but if I needed to move something, it makes it easier. Yes, there's some underlying services that I would have to rewrite and I'm looking at all of those. I've talked to customers that started with Kubernetes somewhere other than Amazon and moved it to Amazon and they said it did make my life easier to be able to do that fundamental, the container piece was easy to move that piece of it but the discussion of multi-cloud gets very convoluted very easily. Most customers, when I talk to them, it's I have an application that I run in a cloud. Sometimes, you know, there's certain, large financials will choose two of everything because that's the way they've always done things for regulation and therefore they might be running the same application mirrored in two different clouds but it is not fall the sun, it is not I wake up and I look at the price of things and deploy it to that environment. It is a little bit tougher. There's data gravity, there's all these other concerns but multi-cloud is just lots of pieces today, more than a comprehensive strategy. The vision that I saw is if multi-cloud was to be a successful strategy, it should be more valuable than the sum of its pieces and I don't see many examples of that yet. What are you seeing when it comes to multi-cloud and how does that serverless discussion fit in there? Yeah, I think your point about data gravity is the most important thing. I mean, honestly, compute is commoditized, right? So whether you're running it in a container and that container runs in Fargate or is orchestrated by Kubernetes or it runs on its own somewhere or something's happening there or it's a fast product and it's running on top of Knative or it's running in a Lambda function or in an Azure function or something like that, compute itself is fairly commoditized, right? And yes, there's wiring that's required for each individual cloud but even if you were gonna move your Kubernetes cluster, like you said, there's rewrites, you have to change the way you do things underneath. So I look at multi-cloud and I think for a large enterprise that has a massive amount of compliance regulations and things like that that they have to deal with, then yeah, maybe that's a strategy that they have to embrace and hopefully they have the money and the tech staff to do that. I think the vast majority of companies are going to find that multi-cloud is gonna be a completely wasteful and useless exercise that essentially is just gonna waste time and money. I mean, it's so hard right now keeping up with everything new that comes out of one cloud, right? Try keeping up with everything new that comes out of three clouds or more. And I think that's something that doesn't make a lot of sense and I don't think you're gonna see this price gouging like we would see with something I probably the wrong term to use but something we would see that sort of lock-in that you would see with Oracle or with Microsoft SQL or some of those things where the licensing became an issue. I don't think you're gonna see that with cloud. And so what I'm interested in though in terms of the term multi-cloud is the fact that for me, multi-cloud really where it would be beneficial or is beneficial is when we start talking about SaaS vendors. And I look at it and I say, look at, Oracle has its own cloud and Google has its own cloud and all these other companies have their own cloud but so does Salesforce when you think about it. So does Twilio. Even though Twilio runs inside AWS, I mean really I'm using that service and the AWS piece of it is abstracted that to me is a third-party service. Stripe is a third-party service. These are multi-cloud structure or SaaS products that I'm using and I'm gonna be integrating with all those different things via APIs like we've done for quite some time now. So to me, this idea of multi-cloud is simply gonna be, it's about interacting with other products using the right service for the right job and if you're duplicating your compute or you're trying to write database services or something like that that you can somehow share with multiple clouds. Again, I don't see there being a huge value except for a very specific group of customers. Yeah, you mentioned a term cloud native earlier and you need to understand, are you truly being cloud native or are you kind of cloud adjacent? Are you leveraging a couple of things but you're really, you haven't taken advantage of the services and the promise of what these cloud options can offer. All right, Jeremy, 2020, we've turned the calendar. Yes. What are you looking at? You're planning, you've got serverless conference, serverless days, Boston, coming up April 6th in Cambridge. So give us a little view as to kind of your viewpoint for the year, the event itself. Sure. Your podcast, you got a lot going on. Yeah, yeah, so my podcast, serverless chats, I talk to people that are in the space and we usually get really, really technical. So if you're a serverless geek or you like that kind of stuff, definitely listen to that, but yeah, but 2020 for me though, this is where I see what has happened to serverless and this goes back to sort of my stop calling everything serverless post was this idea that we keep making serverless harder, right? And so as someone who's a serverless purist, I think at this point, I recognize and it frustrates me that it is so difficult now to even though we're abstracting away running that infrastructure, we still have to be very aware of what pieces of the infrastructure we're using. We still have to set up the SQSQ, we still have to set up EventBridge, we still have to set up the Lambda function and the API gateways and there's services that make it easier for us, right? Like so we can use a serverless framework or the SAM framework or Arc codes or architect framework. I mean, there's a bunch of these different ones that we can use, but the problem is it's still very, very tough to understand how to stitch all this stuff together. So for me, what I think we're gonna see in 2020 and I know there are hints of this serverless framework just launched their components. There's other companies that are doing similar things in this space and that's basically creating, I guess what I would call an abstraction as a service where it's essentially, it's another layer of abstraction on top of the DSLs like Terraform or CloudFormation. And essentially what it is doing is saying, I want to launch an API that does XYZ and that's the outcome that I want. Understanding all the best practices. Am I supposed to use Lambda destinations? Do I use DLQs? Like what should I throttle it at? Like all these different settings and configurations and knobs, even though they say there's not a lot of knobs, there's a lot of knobs that you can turn encapsulating that and being able to share that so that other people can use it, that in and of itself would be very powerful. But where it becomes even more important and I think definitely from an enterprise standpoint is to say, listen, we have a team that is working on these serverless components or abstractions or whatever they are and I want team X to be able to use, I want them to be able to launch an API. Well, you've got security concerns, you've got all kinds of things around compliance, you have what are the vetting process for third-party libraries, all that kind of stuff. If you could say to team X, hey, listen, we've got this component or this piece of, this abstracted piece of code for you that you can take and now you can just launch an API, a serverless API and you don't have to think, you don't have to worry about any of the regulations, you don't have to go to the attorneys, you don't have to do any of that stuff. That is going to be an extremely powerful vehicle for companies to adopt things quickly. So I think if you have teams now that are experimenting with all of these little knobs, that gets very confusing, it gets very frustrating, I read articles all the time that come out and I read through it and I'm like, this is all out of date because things have changed so quickly and so if you have a way that your teams and somebody who stays on top of the learning of this can keep these things up to date, follow the most, the leading practices or the best practices, whatever you want to call them, I think that's going to be a huge, a hugely important step for making it so that teams can adopt serverless more quickly and I don't think the major cloud vendors are doing anything in this space. I mean, I think SAM is a good idea but basically SAM is just sort of a rewrite of the serverless framework, whereas I think that there's a couple of companies who are looking at it now saying, how do we take this, you know, whatever, this 1500 line cloud formation template, how do we boil that down into two or three lines of configuration and then a little bit of business logic because that's where we really want to get to is just we're writing business logic, we're nowhere near that right now. There's still a lot of stuff that has to be done around configuration and so even though it's nice to say, hey, we can just write some business logic and all the infrastructure's handled for us, the infrastructure's handled for us if we configure it correctly. Yeah, it really reminds me, some of the general threads we've been talking about, cloud for a number of years is, remember back in the early days, it was cloud was supposed to be inexpensive and easy to use and of course, in today's world is neither of those things. So serverless needs to follow those threads, love some of those viewpoint, Jeremy. I want to give you the final word, you've got your serverless day, Boston got your podcast, best way to get in touch with you and keep up with all you're doing in 2020. Yeah, so at Jeremy underscore daily on Twitter, I'm pretty active on Twitter and I put all my stuff out there on the serverless chats podcast, you can just find it serverlesschats.com or any of the pod catchers that you use. I also publish a newsletter that basically talks about, talks about what I'm talking about now, every week called off by none, which is collects a bunch of serverless links and gives some IO pine on some of them. So you can go to offbynone.io and find that and then my website is jeremydailydaly.com and I blog and keep up to date on all the kind of stuff that I do with serverless there. Awesome, Jeremy, great content. Thanks so much for joining us on theCUBE, really glad and always love to shine a spotlight here in the Boston area too. Appreciate it, thank you. I'm Stu Miniman, you can find me on the Twitters, I'm just at Stu, STU, of course, thecube.net is where all of our videos will be, will be at so many events in 2020, look for me, look for our co-hosts, reach out to us if there's an event that we should be at and as always, thank you for watching theCUBE.