Red Hat OpenShift Upstream AMA | OpenShift Commons Briefing, 30 Nov 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OCB AMA: What is Quarkus and Why Is It The Ideal Java Runtime for Cloud Natives?

Description

Modern application development is evolving to support new architectures like microservices, containers, and serverless. It’s no secret that traditional Java frameworks are not ideal for creating these types of applications. Instead of learning a new framework for cloud-native applications, you should leverage your Java skills with Quarkus is specifically designed to optimize Java for container environments. Quarkus, in combination with OpenShift, provides an ideal application environment for creating scalable, fast, and lightweight applications. Quarkus significantly increases developer productivity with tooling, pre-built integrations, application services, and more.

Guest Speakers:

Rich Sharples, Application Services Product Management Senior Director
Jeff Beck, Application Services Product Marketing Manager
Daniel Oh, Application Services Technical Marketing Manager

A

And then after I introduce them rich, you can give them deeper, well, hello, everybody and welcome to um openshift commons. It's monday post thanksgiving, so everybody's back and rested, and we thought it would be a great time to um do a deep dive into quarkus java and all sorts of fun things. And so today we have rich sharples, jeff beck and daniel.

A

Oh, who are going to kick us off with a quercus talk, so I'm gonna, let rich introduce everybody a little bit deeper and tell us what we're gonna do and we'll have q a at the end after everybody gives their spiel and demos, and I'm really looking forward to this. So thank you. Everybody for coming.

B

Okay, thanks dan, so um my name is rich sharples, I'm with uh jeff and daniel here, uh we're all part of the uh the application services business unit previously middleware and today we're going to talk about java and caucus, and you know really where java is headed.

B

um I will start off talking about some of the uh the kind of market forces and the reason for or caucus and um then jeff's going to talk uh specifically about uh the benefits caucus on openshift and what's in the latest uh version of caucus and daniel will do uh do a demo um which will be pretty exciting so stick around for that and then we'll follow up with uh with q a and we'll be here to take your questions so without further ado, let's uh jump into the next.

B

um So, actually go straight to the next one. I mean yeah, one of the questions uh I've had you can stick with the graph the chart actually over the years is uh yeah is java still relevant, and my answer for the last I don't know, 20 plus years has been yeah. It's still highly relevant.

B

uh It's really the dominant language used in the enterprise, whether you're a bank or an airline building complex distributed. You know transactional systems in all likelihood you're using java for all some of that, uh it's been one of the most popular languages uh in the industry for last 20 plus years according to toby.

B

In fact, this month was the first time java had dropped out of the top two uh python uh took the number two spot, which is not surprising. Python is going through a real renaissance right now, with ai and ml.

B

um There was something like you know, between 7 and 12 million java developers, out of the 24 plus million professional developers worldwide. So a big chunk of existing developers uh use java as their primary or secondary language, adopt open jdk, which is the place where most people download java gets gets about. A hundred thousand downloads every single day, and he just hit the 200 million download mark after he also so uh java, is still very interesting. It may be uh quite old, it's 25 years old.

B

I think python is a little older, um but languages tend to stick around for a long time. You know when, when you have a couple of trillion lines of code in java or cobalt or c or c, plus, plus um that that's sticks around for a long long time, people who have a big investment in skills and systems and technologies on a particular language or platform uh creates a huge momentum.

B

So what I'd say is java's likely to be very relevant for the next decade and we'll likely be around for another 20 30 years next slide. Please.

B

Now you know java was not designed for the systems that we build. Today uh it was designed to run on big machines- big sun buyer. You know.

C

B

With you know, hundreds of cpus hundreds of gigabytes of ram, it was really designed for big throughput systems, uh irrespective of the amount of memory that would be used. It would just use all of the resources on a system uh designed for these long-running systems that ran banks and airlines, and it had rich, dynamic behavior. So if you need to load a new application into jboss or into webster, you can load a new conversion. Application change, classes dynamically, so lots of really really clever stuff.

B

That was really the environment for which this java was was not originally designed. It was actually designed for small devices, but that that's where it really found its uh it's it's footprint in the 90s and 2000s. Now, when.

C

We look at systems today.

B

um We we solve the throughput by scaling out, um and we typically do that by scaling out containers managed by kubernetes. That seems to become the uh rapidly becoming the detractor standard for modern systems. um These systems are ephemeral and immutable.

B

When we want to upgrade a software component, we just blow the whole container away and replace it with a new, updated one. When I'm running as a function, my code may run once and then it's gone.

B

um I have no opportunity for doing some of the the very very cool uh profile and optimization that java was designed for and in these new systems. You know footprints memory footprints when we're talking decomposing our existing applications into tens or hundreds of microservices memory. Overhead matters, um the overhead of the jvm, really does matter and startup performance methods, as well as, if I'm bringing uh applications or functions into into being just a service single request.

B

D

C

B

Design center today, for which java was never really designed for so next select things, and when we look at people building applications in container platforms, um it's pretty pretty clear that you can get much higher density um with something like goliath or node.js, while they may be not as good for some things in terms of their efficiency and agility, they do have an advantage.

B

um So yeah next slide when you, when you survey something like um uh next slide. Please if you survey something like uh the um the new relics, um amazon, lambda um um stats uh java was pretty late to the party there. Yeah no gs for the longest time was really the primary language runtime for lambda, followed by python, and even now it still lags because these languages just much more efficient running in those environments.

B

So caucus is really designed to solve this problem. How do we make java great again in these new environments that next swipe is so what we focus on is is cloud native architectures running in containers running in kubernetes running in those those kubernetes managed systems um for developing microservices or functions or serverless applications. That's really the focus of caucus um if you're riding large monolithic, long-running applications, something like jboss eap, is likely a much better option for you, um so caucus is focused on cloud-native, microservices and service next slide. Please.

B

um So just kind of talk through some of the main benefits of caucus.

B

We've talked a lot about the performance and the scalability, but um what we found is that, while people initially come to caucus um for the footprints and the performance and startup time, what I actually stay for is is that the developer productivity, when you're, when you're developing in a servless connective or a kubernetes environment um caucus, really does have a very very nice developer, um especially for that inner loop development, where I'm deploying and testing and changing pretty quickly and not having to go through a full kind of build and redeploy um so very, very convenient for the developers and that's important, because developers are still one of the largest costs of any software project.

B

Next slide, please.

B

The next benefit is really around the the uh the memory footprint. So caucus can run in a couple of different modes. uh You can run in a traditional jvm, open jdk, for example, so you can still get some of the benefits of of hot spot kind of performance, profiling and long-term improvements in execution, or you can run compile down the native using say, grail vm.

B

So just a comparison of the uh the the resonant set size memory. That's the true memory as you're running the traditional cloud native stack on the right hand: side! um If you run, if you qualify your application and run within jvm you download about half, and then you know, if you compile to native as well, the memory footprint is significantly reduced.

B

So when you're running hundreds of microservices this, this really does add up the next slide. Please so slightly more complex application. So you know a restful endpoint with access to a database. Again we're measuring memory here again again, just going from traditional jvm traditional stack versus caucus running on open jdk, we're not quite half the size, but still a significant reduction there. But then, if you compile down a native and remove all that dead code and uh your libraries and overheads you don't need, um we can significantly shrink yeah.

B

It's almost a tenth of the original native stack in terms of memory, again huge savings to be realized there in a a microservices environment where you have lots of uh small services running next slide. Please.

B

um So so now, let's kind of talk about startup time, and by that we mean you know the time it takes for you to deploy and then actually be able to respond to the first request, whether that's be a rest call or for a http. You know, request response and that's important. If you have pods sitting around unable to service requests, they're they're doing nothing for you, they're essentially dead to you until they can get to a point in time where they can service requests so caucus.

B

You know, I guess again compared to a traditional cloud native stack versus uh running in a jvm running in uh native mode. The startup speed is is just phenomenal. um That's because lots of magic happens. You know we convert a lot of startup code um to kind of compiled code of compilation, time versus run through when you first run, so we just do some real magic behind the scenes, but then getting down to a native image really does uh uh significantly change the response times.

B

The first startup and that's uh really important for event, driven serverless functions and slightly.

B

The um benefit 3 is really related to developer experience. Again, um caucus allows you to develop uh in a familiar imperative style that most job developers would be used to or if they're, building your asynchronous reactive systems, maybe using kafka and they're handling things.

B

You know, as events with reactive programming would be more more natural and more kind of a natural expression that logic um we support a true reactive programming model as well, um true, truly asynchronous, truly reactive, and that that's a huge um bonus again. If you're developing your high scale, high scaling microservices, maybe an endpoint for an api or for a dream, but with kafka.

B

Developing um using the reactive program model will not only be a. You know, provide more readable, more maintainable code, but it's likely going to be a lot more performance and scalable as well. So that's a that's a pretty big deal, um but the good news is you don't have to. um If you want to tradition your traditional imperative style there we handle lunges the next slide. Please.

B

And then yeah, finally, the the the real benefits of uh caucus are: it's still java, so with um some or or no modification, most of your existing technologies, whether it be camel or rest easy or keep cloak that can still be used within your applications. You don't have to learn a new language, a new framework.

B

um You can bring your knowledge and maybe even some code with you as you move into this kind of cloud-native architecture with caucus. That's pretty huge and with that I will hand over to jeff.

C

All right, thank you, rich hello, everyone. My name is jeff beck, I'm a product, marketing manager on the application services team, and I want to pick up where rich left off and talk more about quarkus uh in the context of openshift and kubernetes, and really you're going to see some repeating patterns. So I don't want to dig into it too much, but why would you be looking at using quarkx for kubernetes native development?

C

We have the tagline for quarkus, where it's supersonic subatomic java, and it really drives home the fact that rich just showed you the performance um and how it's able to use or use the memory and read cloud resources efficiently, while also being uh fast and and first from a startup perspective. So what does that mean? It really translates to cost savings.

C

So we talked about cloud efficiency, uh serverless deployments, but also we talked about developer productivity, so not having to retrain your existing developers in a new language, allowing them to use what they know and what they're familiar with, and you know, essentially that translates into faster time to market. So all these tools allow a developer to respond more quickly to a changing customer need changing.

C

uh You know, marketplace demand, or even sometimes internal pressures uh within an organization, so there's a low learning curve and allows organizations to be competitive, and you know from a kubernetes native perspective.

C

Quarkus was built from the ground up uh for kubernetes, and so there is a integrated tooling, uh like the dev console on openshift, as well as code ready workspaces, as well as uh the ability to use the extension ecosystem that that rich mentioned, where developers can leverage the extensive ecosystem uh to to do things to, I guess, focus on development and not have to worry about the underlying infrastructure, for example, using the open shift or kubernetes extension for quarkus that allows them to.

C

You know, use a single maven command line to deploy that application and create the underlying resources and kubernetes uh to support that you know. So what does it mean to you know for kubernetes native development with quarkus, like we mentioned before it starts with the solid foundation, uh so the java that they're familiar with and they know but applied to a modern architecture.

C

So cloud native kubernetes native uh deployments, uh stunning performance, like we talked about optimized, especially when you talk about the native quarkus, is very compelling when compared to a traditional framework when we're looking at you know running on the jvm, but if the application has the ability or the need to be compiled down to native uh the performance becomes even more important, especially when we're talking about serverless.

C

I talked a little bit about this tool chain. Really this uh you know being able to integrate more tightly with uh openshift and kubernetes to utilize things like that developer, console, um there's, uh quarkus extensions for all the popular ides. Where you can uh generate the scaffolding of a project. You can easily uh include extensions in the palm xml for using maven uh really.

C

So all these tools in live reloading, which I'm sure daniel is going to show you is, is one of the one of the major features that developers like to make them more productive, and you know rich talked a lot about the community.

C

You know first being focused on you know from java and everyone's experience with java, but the extension ecosystem is really starting to become an important part of developer productivities so, like I mentioned, allowing the developers to focus on their business knowledge and their expertise and not have to worry about underlying infrastructure or how to integrate with applications or underlying databases.

C

So you know: we've been talking a little bit about you know the developer and and obviously openshift and kubernetes are more uh traditionally focused on the operations team, and so what's interesting, now is you're starting to see these two worlds collide with the operations and the developers. um I guess clyde is probably not the right, the right analogy, because you know they've been working so closely together.

C

But what is interesting is that now we're we're looking at a combination of corcus and openshift as having a common uh development and operations environment for this cloud native development, there are a number of customers using uh quarkx in production and you'll, see just a few of the quotes here.

C

uh Lufthansa was able to run three times denser deployments without sacrificing availability and response times, so they had an existing portfolio of applications and they were they were able to take those do more with what they have right, so they were able to run applications quicker and they were uh more efficient with the cloud resources.

C

I like the one from talk desk here was they were looking at uh more traditional frameworks like spring and they weren't satisfying their needs from their. You know, container-based applications, and so when they they looked to adopt corcus, they were productive from day one.

C

The idea once again that it's leveraging the existing expertise, not some developers, love to be on the cutting edge and and use the uh different frameworks for the you know for the needs of their applications, while others really like to lean on their expertise and what they know over the last you know could be 10 25 years.

C

We also, you know, we talked a lot about performance so where traditional frameworks may not be suited. For you know, modern cloud native architectures uh quarkus really uh delivers on that front.

C

So when when, and why do we use quarkus a number of different things? First, you know net new from a greenfield perspective. It just makes sense if you have the ability to you're starting and creating a nuna application, and you know java, it's it's a it's a great place to start so um same things. We talked about low memory. Footprint fast startup times, um for so creating microservices using quarkus is a is a great option if you're looking at how to modernize existing java applications.

C

So this this world of uh taking the monolith and- and you know, turning into either microservices or bolting on uh functionality. Quarkx is another great choice.

C

From that perspective, uh we've talked about serverless, so you know scaling up or down in an extremely fast manner and to riches point being able to not only just scale up and scale down, but to be able to handle a transaction or a process immediately is a great use case for corcus and then, as we mentioned, we're really trying to focus on being able to support event driven architectures, so these asynchronous reactive event loops that you know make it easy to create these reactive applications.

C

So just a just a few links here or some names here to talk about who's using corkus in production. There is a uh on the quarkus.io website, there's a section specifically geared towards customer success stories, and these are really interesting to read because they come from a bunch of different perspectives. Of some are business driven. Some are technology driven, and it's really interesting to see how these these users, you know how did they find out about corkis and what? What?

C

What is a real experience, a lot of stuff around uh folks that are using traditional frameworks like spring when they were evaluating? You know our spring applications aren't performant in this new containerized, uh kubernetes or openshift world. You know we're looking at moving to node.js or go and that you know we were thinking about. How do we retrain everyone? When corkus came along we you know we saw that it was a ability to use our existing java skills. So I really encourage you if you're interested, take a look at some of those success stories.

C

They're they're very interesting to look at individual customers.

C

You know I wanted to talk. You know, lastly, about I guess the great trifecta of reasons to start developing or continue developing um your job applications on on openshift and kubernetes, so one you've got on the one side. You've got the the proven leader in java, with our experience from a red hat perspective, and then we've got this de facto standard, for um you know, contain containerized solutions and microservices with kubernetes, and now you've got this growing community with corcus corcus was introduced.

C

I guess it's going to be going on almost two years, starting next year in january or february, and it has grown exponentially with a a large number of contributors. This extension ecosystem that I was talking about um quick cadence for release, and it's really it's really amazing, to see how quickly it has evolved. So when you look at these three things, you know, I don't know if certain people are risk averse or not risk averse, but these three things tell me that you've got a proven leader, a proven platform and a proven community.

C

That's um it makes it less risky to start taking a look at it, especially from a production workload perspective.

C

So this is kind of a little bit more background of on why you would use and how you use quercus on on openshift, I wanted to talk a little bit about specifically about the red hat, build of quarkx release. um So what's new in this in this release versus uh one three, so one three was was developed or or launched in uh it was about april of 2020 and in october time frame we release 1.7, so you'll notice, these numbers 1.3 and 1.7.

C

They are directly tied to the community project, build number so to eliminate confusion. So if you're, if you're, looking at the red hat build of corcus, obviously it's based on the quarkus community, build of 1.7, and that was purposely done. So there's some consistency there not trying to figure out. um You know supportive features and functions so with 1.7, which was released in october.

C

The there's a couple of things that really stand out and, first and foremost it's the native compilation with mandrel 1.3, did not support native compilation only supported on the jvm, so it was important for red hat and for us to uh obviously quickly come around and support native compilation and you've seen the numbers speak for themselves and why um why it's compelling? So that was one of the major features. Another one is spring compatibility, so corcus has has, can and continues to have, uh support for spring developers.

C

uh There are a lot of spring developers who are looking to to move to corcus, and so it just makes sense that we have this compatibility layer that allows that transition to be a little more seamless. So with one seven we've got config cache and scheduling, I believe library supported. So that's in addition to data web. I remember off the top.

C

I had now off to take a look, but what the point is that it's really helping spring developers uh transition into corcus and there's some additional tooling as well, for serverless so being able to serverless eventing. Excuse me um where it allows you to easily deploy your applications onto openshift using an extension there's a couple of things that were in tech preview.

C

One is the cute templating engine, as well as continuing to support this reactive uh programming, with uh support for mysql and db2, and then um on new supporting new new certifications, including openshift45, and, like we mentioned, also supporting serverless serving with eventing soon to follow, so that's be on the lookout over the next quarter, two for supporting eventing as well, and then continued integration with other red hat products like data grid and ssl, and I guess most importantly, also extending now the life cycle support now to march 2021.

C

You know you probably were noticed that we said native compilation with mandrel and so with native compilation. So why is native compilation support? We mentioned this, the the low memory, consumption, the lower cpu usage and then it starts and is able to you know, process transactions uh in milliseconds.

C

So, oh sorry about that up up until this point in uh version 1.3 we were support or what was included was uh was graw vm, so with 1.7. Now, what we're doing is we're actually leveraging mandrel uh to support the native compilation. uh So you know so. Mandrel is a downstream distribution of grawl vm.

C

So it's a you know, direct downstream of growl, so growl is a a large open source project that has a number of contributors and a number of different features and functions, and so red hat determined that it was easier to support a downhill, a downstream distribution of it to support customers, make sure it's secure and then be able to productize the version. So so mandrel offers the same features and capabilities that that growl has including things like debugging and monitoring.

C

It's based completely on open jdk, so uh grau vm has been forked and is often updated and changed.

C

Mandrel is based completely on open jdk, and this is important because we want to make sure that there is no need to pay oracle for a subscription um for growl vm, there's a different licensing available for growl vm, but this just makes it cleaner for us to be able to uh make sure that you don't have any additional support costs for gravia and one last thing I want to lead you with. Is you know? How do I get started? There's a couple of great ways to get started.

C

You can use uh the ide extensions and there's also uh code.corkus.com and code.corks.redhat.com, which is very similar to the spring initializers, where it allows you to choose the extensions you want, deploy an application, the basic framework of an application, a hello world.

C

The minor differences between code.corkus and code.corks.red hat is that the red hat version makes it easy to see which extensions are supported versus tech, preview and unsupported. So if you're interested in getting started, we recommend there that's a great place to start and with that, let's stop talking. Let's actually see some demos daniel. Do you want to take it from here.

D

Thank you, jeff yeah. Let me share my screen. First.

D

Okay, hopefully everybody can see my screen okay, so my name is danielle. I'm technical marketing major. So I'm going to showcase today around a multiple approach to a devil apps in terms of how to create your quarkxp application and then from scratch. And then some serverless application with the creator function uh also some benefit for ops team with the hamster. Okay, here's my id2 vs code, but actually you can use any id tool like intellij and eclipse or whatever you need and then uh the quarkx tool. uh You can actually install your paper id tool.

D

So you can start the cortex project created from your id tools, so you can usually uh use the maven or gradle to package your application.

D

So I'm going to create a new progress application on my id tool and one of the benefits of a parkour style application. uh You can actually uh uh create some sample application here and then, as you can see, uh the data automatically generated some sample application. For example, the quarkx uh provide uh the respo api by default and the full support cdib.

D

So you can have the sample application how to expose a restful api based on focus application. So, first step we're going to start this application on my local machine to make sure everything is okay, so I'm going to open my terminal window and then I'm going to use the maven command line as usual for developer standpoint, so maybe parker's the demo. This is one of the beauty of the purpose of the reach and jeff already mentioned the rear.

D

The quarters provide the line coding capability, which means whenever you change your code, uh you don't need to recompile, rebuild and re-deploy your application on your local machine because of the line coding capability activated by default when you're using corpus demo. So what exactly that means for developer?

D

So let's go to open new uh web browser and go to localhost 8080 and then uh go to rest, easy hello. So we have a return and hello. So whenever, let's say okay, I need to change my code like, for example, very simple. uh Let's change it to uh the written code here we come uh appreciate.

D

Commas, like that, and I just saved my file and then go back to my web browser. I try to reload my application and I immediately get a new return result in the meantime. Behind the scene, progress automatically detected, which java class is just change and stop and recompile review and the rerun.

D

This application automatically suggests a huge benefit for developer standpoint and in the end, uh this feature increases development, productivity for your daily work and then I'm going to stop at that mode here and then I'm going to package this application before I deploy this application into kubernetes cluster. So I'm going to use the operation container platform built on kubernetes for today demo. So in order to do that, I'm going to clean, uh you know, package this application and I'm going to skip unit test because I changed the written code.

D

So uh after this packaging application, you've got uh the rollable phene jar. As always, you can run the application using a java command line. So here's my activity, monitor on my local machine, just to make sure uh take a look at that.

D

uh How many jvm I'm using right now and as you can see, I'm going to zoom in a little bit, so I have actually full jbm uh running on my machine and the smallest one 41 and the second consumer is 185 just to remember this number and go back to id2 and I just run my application of the target 3 and the servers and runner.

D

So once you run your application and you can see you just need to here- 0.6 second, to start off there we go. So this is a pretty fast than any other cloud native javascript.

D

So the spring will maybe a couple of seconds to start up the application, even simple uh restful api, but it's almost four times faster than input or any other java application framework. And then you can also have the same functionality when you go to the application.

D

Okay, you got us the same result here and then go to memory uh utilization, and now we can see here, 105 megabytes as your memory for free free, uh small memory properly to start up and run our quarkx application and let's stop this application, and then we need to uh uh the packages application.

D

Just before that, we just need to change the application to serverless function, because the a lot of customers and a lot of enterprise companies are considering how to evolve existing microsoft, application to serverless or function, because the memory of footprint and the resource consumption is a really big concern uh to use uh cloud infrastructure, specifically optical v19. So memory fulfilling or research management is a critical issue for every single company. So the qualcomm should luckily uh provide a good feature to change your easy microsoft. Application to serverless function very easily.

D

In order to do that, we needed to add one extension from quarter side. So it's a funky extension. The funky extension is to provide the standable java portable api and you can deploy. You can actually develop the microsoft application as a function and then you could deploy that function with a multiple serverless and fast platform such as amazon render a google function, azure function and also uh you know, on-prem kubernetes uh serving equivalence event as well.

D

So once I added the quarkus funky extension- and I just need to use that part of uh func annotation here uh instead of the existing uh request response uh annotation like a get and pulse method. So I just add okay annotation here and then let's try it one more time.

D

Annotation here now this uh this uh and I just needed to delete unnecessary packaging. As you can see, I even tried to uh change this microsoft application to simple function, but just the last code and pretty simple, and then I need to uh make sure uh the function name. Here I got parker's, you know func. I can export name and hello so because uh the function name by default is the same name of your method.

D

Name like hello, but if you can actually specify a specific function, name with the, for example uh like a serverless, and then you could change that serverless. So this is the way uh how to change your using microsoft application to function now we're gonna uh package, this application with the native compilation. The jeff already mentioned here- that one of the huge benefit of a quarkx application, specifically uh serverless application runtime.

D

So in order to that, I'm going to using the same maven command line here, maybe clean package, I'm going to skip the unit test here and then I needed to use the native com profile to packaging this application.

D

So it takes a little bit longer than a usual maybelline packaging, because in the meantime we created thin jar just like we made with packaging previously, and then we need to uh package it one more time at executable by using uh based on native compilation.

D

Well, when you use when you use the red hat, build of the progress we're going to use the name, a mandrel of builder, to build this application and an executable file, it takes a little bit of humanity uh to complete, uh create a secure file.

D

So in the meantime, uh let's take a look at that my uh openshift cluster. So here is my opportunity cluster, as you can see, so here's your new project, serverless demo and then there is no uh resource here, because I don't try to deploy any uh application at this moment and go back to my ide tool. It's almost there and once we're done, we can deploy this application to uh or push the container platform using our uh the quarkus uh uh opposition extension. This is one of the benefit of a quarkx.

D

You can actually deploy uh corcus application from local machine to operation container platform remotely okay, it's just done and then, when you go to target creatory you can see here is a runnable executable file. So how do we execute this log of a file?

D

You don't need to java command line any longer, just run that file just like experiments like gold, python, etcetera, so uh coverage and run, and now we just need to uh 17 milliseconds to start off this application exactly the same functionality, but previously we just needed 0.6 seconds to start up this focus application, but now we have just 17 milliseconds to start the huge impact, your response time emma and also when you go into uh memory property here. So previously we have a 110 megabyte as a memory for free.

D

Let's take a look at the new here. So here is the 7.2 megabyte. It's almost 25 uh redstone memory footprint uh previously uh previously the thin jar application and then I'll go to a local environment and I'll try to reload, and then uh we just change the uh endpoint like a suburb list and that's easy.

D

Okay, we got the same, welcome opportunity comments. So the same application features and functionality, but the developer uh can choose the two different packaging technology with the hotspot on open jdk, but also they have enabled corporation. So it depends on what kind of workload you need to deal with uh on kubernetes cluster.

D

So now I'm going to start a native corporation and then in order to deploy opportunity platform. I need to add a few uh configuration here. So let's try to uh deploy this application using maven command line, maybelline clean the packaging and I'm going to screen unit test one more time and adjust it and then behind the scene. We did this.

D

uh I just forgot one more thing: I just needed to add one extension like a print shift, so focus on which extension allows you to deploy your app percocet application to container platform, and actually I needed to make sure. um Where am I uh in my projects over this demo and then I just try to make one package one more time.

D

So, as you can see in the in this computation, I'm going to make it true to contain a bill and then I'm going to target environment the knee, which means this extension will deploy our caucus application as k, never service.

D

You can also deploy uh traditional application like uh to kubernetes operation, is a 24 7 running power like a staple application, but in this case we're going to produce application as the function uh using opposite opposition to the serverless capability, and also uh there are command sd pro 2 in a exposed option to exploit true means opposition to automatically create a router to uh for developer to access from external.

D

So it takes a little bit longer because behind the scene we need to package the application at the finja and then deploy opportunity container platform using a source to image processor.

D

So when you go to uh the operation container platform, uh it will be show up uh our quarkx application in a few minutes, I think, and in the meantime maybe uh we can try to another different way to create the function a very quickly and then easily. Okay, we just done and go back to here now we gotta have a new uh focus application as a server lens and go to view row and you can find the application running in jvm.

D

So actually, when you add more uh another configuration, if you want to deploy this application as the uh name name corporation article, you can use to add a name compilation build through and then using network imagery like a mandrel, the java mentioned a rear and then, once again you could deploy this application using maven command line. uh You have uh native application rather than jvm, but it takes a little bit longer than longer than the usual case.

D

So I'm going to skip that thing uh in this demo and once you have this application and go to url and you have a landing page and then uh separates, and you got us the same result here, so this application, uh serverless application will be scaled down to zero. um Maybe next!

D

uh Second, because there is a default computation uh time to scale to zero uh in the kubernetes k net survey and then and after that, when you invoke this application using a restful api, just like we did in a web browser, and this application will be scaled up automatically in a few seconds, just like a serverless behavior okay. So this is the way how to get started and how to develop your purpose application from scratch. But sometimes you don't need to go through the every single detail about this application.

D

So I'm going to show you another cool thing here, so the k-network service command line, uh kn uh create canyon stuff, but we're going to add a new feature in race opportunity for six with the opposition stabilized opera 111 version. I still got preview, but it's pretty interesting for many developer to create a function uh project, for example uh kn func, and so let's try to change the project first uh here in the in. In the meantime, you see our corporate application scaled down to zero.

D

Already we changed the our purpose: application at the corpus function there. You know application department at this moment in km func and then create uh parker's found the function name and uh the runtime is quarters.

D

So when you create that function and we have a new project that called the quarkus funk and then you could find that in all the maven project and then the next step, you have just deployed this application using external registry, for example, I'm going to use my create.io, my external, raise, 3 and and I'm gonna do uh specify the namespace for corporate spawn my opportunity uh project name, I'm going to check my logo during the build.

D

So first of all, I'm going to use the quarkx jvm build.ng and the runtime usually, and then we're going to use the buildpath to packaging this application and deploy the openshift container platform. In the meantime, the this uh k n service fun command line uh to try to uh make one packaging just like we did and then just we just build success. And then we just push this uh container image into external registry and almost done and then the last step.

D

The k-net service uh command line tried to speed up our the the quakers application to suffer less. In our opposition to project, let's go back to opposite container console, and now this application is almost running and just running and as you can see, the funky uh exchange are already there and the application is running and click on the application.

D

So we got us the end point and I needed to actually uh define uh the uh the endpoint uh we tested copy from the endpoint and echo uh for example, uh message.

D

My name is daniel.

D

And the endpoint here there we go so uh what else? Oh okay! There are two double quotations here here: let's try to almost find okay, we gotta also see uh the result here and this application will be scaled to danger automatically just like a serverless behavior. So I have a showcase already two, a different operative dev apps. I mean more of a custom application side because the we have already here awesome infrastructure, kubernetes container platform will manage your application on time.

D

So this is a totally different way to approach a quarkx application development for serverless and function. I'm going to showcase one more thing: it's available real, quick! So luckily the race opposite, the container platform provide the uh hem chart. So this is another way to how to manage your corpus application with ham chart. It may be really more interesting from uh office people rather than developer, but because the operation container platform and quarkx uh can handle a depth and ops as well so go to uh ad.

D

You can actually add a lot of things from here: developer, console and click on hem chart and find the partners.

D

So there are one quarter, cam chart, you can install crockett's hamstring immediately and then I'm going to add a simple uh the build application from or one of the my uh give repository actually- and I just install this application. There are some missing thing: okay,.

D

Oh yeah, maybe this already came to okay, so this is the ham chart. I uninstall.

D

And then go back to add in ham, chart and a quarters and install hanger and then install, and you have already the quarkx application it in the meantime, when you click on the build log, the first step we're going to clone the external git repository, uh we'll basically implement quarkx application and download, and then packaging build and then deploy the operation container platform. It's very easy, and once you install this quarkx application, you can change the git source code or deployment config or some condition or build water security stuff.

D

And then you can also define all specifications on him chart, which means that hem chart allows you managing your quickest application, upgrade update or some patching etc to go to topology view. uh It takes a couple of seconds and you go to here so in the meantime, you could maybe can find the image you pull back off because the uh when you uh proper the bill is complete uh and you can still see that thing when you go to uh ham child release, no it! This is the notification.

D

So maybe you can find the error image pull or image back off until your build is complete. So go to here and then go to build learning, and it's almost done. It takes a little bit longer because if the download necessary maple packaging, so in the meantime, let's go check it out uh the quarkx.

D

The font application here then go to top quality view. Okay, this is already scaled down to zero. So let's try to uh call one more time with the different uh messaging values of my friend name, yeah back and then we try to call and then, in the meantime, your the quarterback application automatically start up uh just like a serverless behavior, and then it takes uh just a few seconds to start up.

D

If you want, if you use quark's name corporation, it's a pretty much if it stands to start up, maybe just one second to start up and go back to our command line. You know you gotta search the correct result. So this is a really cool thing uh to design software as a function with a purpose application with the jvm, but also navy configuration stuff.

D

Okay, so go to focus 10, and then this application almost done. We are running over time, so so just a wrap up of okay just down and then go to rom application uh just make sure the quarkx application is running, go to view log and the application is up just now and go back to lower, and now we have the ciud and the application. As you can see, uh this application based on the retail, build or quarters the race button 175, and then he here is just a simple cloud.

D

When you want you can create any application of the task like I'm going to create a demo and then maybe a writer write, a blog and actually you can go and uh delete the application. So so what what do we run today from this demo? You can actually start the quarkx application development and manage it from local environment to deploy the application as a serverless, but also traditional microsoft, application from scratch, while using kn, func cli command line, but also you can use the hem chart for more operation team perspective to manage your purpose application.

D

So that's all! That's all. I have I'm going to hand it over back to jeff.

C

All right, thank you, daniel um I'm gonna, we're gonna finish off here with um just some additional resources. So, as you are sorry about that one. Second, um if you want to hear more about what we've talked about, there's a there's, a number of places you can go. We typically suggest people go to corcus.io where you've got information on how to start developing applications. All the guides and extensions blogs about some of the customer stories we talked about, so that's always a great uh place to go.

C

um There is also an interesting idc report, lab validation report that was done recently that compares spring in a number of lab validation, study scenarios compared to traditional java frameworks. So that's a that's a great uh asset. There redhat.com has a bunch of information uh on uh on quarkx as well, and then red developers.redhat.com has some interactive tutorials and those types of things, so those are great places to start.

C

So it looks like there's a. There was a couple questions that that came in one question was which red hat skus contained support for corcus and the there was a recent announcement two weeks ago now. I believe that quarkus is now fully supported uh on existing openshift skus, so corkus is now included. Support is now included, as well as where it was traditionally in the middleware and the runtime skus as well. So it's now available in all three of those locations uh from a skew perspective, and there was another question around.

C

We talked a lot about serverless, and so there was a question around a funky. So funky is a extension on the community side that allows you to create a common, a portable api for serverless applications, whether you're deploying on amazon, lambda or azure functions or openshift, k native and so there's a lot of excitement and energy around this funky extension. um So it is supported via the community.

C

The funky extension will most likely, hopefully be supported in the next version of the red hat, build a corcus, but at this point it is a community supported um extension daniel anything you want to add to that. I don't know if my funky description was great yeah.

D

Yeah, that's a good enough and just one more thing in a more developer standpoint and you don't need to change any uh application code itself when you deploy uh the same funky application to multiple uh serverless platform. Just you need to add some specific configuration on your application properly file, and then you can deploy the same application to serverless platform like amazon, it's all about amazon, lambda to microsoft and google and then openshift serverless. This is a huge benefit.

D

It's a portable function, api.

C

We're running out of time, but there was one other question I thought was really interesting: it came from a user that was saying that they only have about less than five services, so would they recommend using quarkus and uh and which mode would they recommend it and uh looks like rich uh sharples was adding some information around?

C

It's obviously tough to tell based upon what those services are doing, but in general, um like we talked about before the perf, the performance and efficiencies you gain are the same, whether it's five services or 500 services, uh the best part, is if you are an existing java developer. The learning curve is little to none.

C

If you're a spring boot developer, we've got the spring compatibility layer. So um really there is no. There is no right number of micro services to start using quarkus. uh One is just as good as a thousand um looks like we're at the top of the hour. So I think what we're going to do is we're going to let everyone go. We we appreciate everyone joining this session and I'll turn it back over to diane.

A

All right: well, we really appreciate all the three of you joining us today um for the tour de force, um lovely demo daniel. Thank you very very much um and it was we need to have you guys back more. I think I think this will be an ongoing conversation and please, if you folks, have questions reach out to them the qarcus.io site. The community is just awesome and they're doing all kinds of fun stuff, so you can get in touch with them there and all these other wonderful additional resources.

A

So we really look forward to hearing your war stories and your successes. So if you are using quarkus in production, I'd love to hear from you and get your feedback and give you the podium so definitely reach out and tell us what you're doing with corcus these days. um We're we're happy to to share the podium and let other people demo a cool neat thing so take care, and thanks, jeff and rich and daniel always a pleasure. So I'm looking forward to having you back take care guys.

C

A

A