Deploy Friday

Azul, the JVM, and the JDK

Simon Ritter and Geertjan Wielenga are the brains behind Azul, a company focused on providing the best Java Development Kit (JDK) and Java Virtual Machine (JVM) possible. It’s also the biggest Java-only company in the world. As Simon says, “Azul is all about Java.”

Azul has two products:

• Zing — A high-performance JVM. As Simon describes it, “Zing is where we've taken open JDK, we've taken Hotspot, replaced certain parts of it, and then make it into a high-performance version.”

Azul focuses on high performance with Zing

The JVM is very powerful, capable of high performance while also scaling well for hundreds, thousands, or millions of users. But Simon and Geertjan wanted to improve performance even further with Zing. They focused on two main areas: garbage collection and just-in-time (or JIT) compilation.

With Zing, the garbage collection algorithm is written to allow garbage collection concurrently with the application threads. Simon explains the benefit of this. “From a garbage collection perspective, we eliminate pretty much all the latency that's there, and it's very beneficial to a lot of applications.” 

For Zing’s JIT compilation, Simon and Geertjan replaced the C to JIT compiler in the JVM with one called Falcon (based on another open source project). Simon says, “By using Falcon, we can actually get better throughput in terms of number of transactions per second through our JVM.”

Azul’s open source contributions

Simon and Geertjan believe in giving back to the open source community. One way they give back is with Foojay (Friends of Open JDK), an integration platform for Java reference material. Or as Simon calls it, “a Wikipedia for Java knowledge.”  Foojay has a Slack channel, Twitter, blog, and events calendar where everyday Java users can connect and share tips and tricks.

Azul has also produced another open source tool called J Hiccup, which monitors the latency effect of the JVM and the operations system hardware on an application. Simon says, “J Hiccup compares the impact of the JVM latency on your application, both with and without Zing. When you show the Zing latency graph, it's just a flat line at the bottom — no spikes. J Hiccup is quite an effective way of proving that we can do what we say we do.”

Try Zing or Zulu, or connect with other Java users on Foojay.

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Defining Fission

Fission is an edge app publishing platform for front-end developers. As Boris Mann, one of its creators, defines it, “Fission is a stack of tools, identity files, and data storage that gives developers everything that's needed to run a full web app using only the front end.”

Out of the box, Fission has some key benefits.

• Runs anywhere: server, phone, laptop, or a mixture

Fission offers great developer and user experiences

Developer-friendly features include:

• DNS and SSL

For users, Fission offers:

• Private and encrypted user data

Fission and identity authorization features

The typical OAuth authentication model uses access control lists (ACL), which have some downsides. The rules get complex quickly, and all authorization requests have to go through a central server, a potential bottleneck which can slow things down. As Boris puts it, “It’s like handing someone your house keys, and now they have access to your whole house.”

Fission works differently. It’s a distributed authentication system built on JSON web tokens. Users have cryptographically signed certificates describing exactly what a given user is allowed to do. “In this model,” Boris explains, “you can delegate permissions to an app. And that app, in turn, can use a bunch of other services directly and sub-delegate things.”

Fission bridges a gap

Brooklyn Zelenka, the other creator behind Fission, concludes, “We've been trying to have these distributed, decentralized systems forever. But because there's this UX challenge, they get less adopted. With Fission, we're very much trying to bridge that gap. We want to make things usable and easy while still giving people as much autonomy and control as we possibly can.”

Try Fission today.

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Reducing risk when deploying code is a huge priority for many DevOps teams, and incidentally, the reason behind our podcast's name, Deploy Friday. Today, we speak with experts Jackie Balzer and Maryann Bell about best practices for risk reduction during code deployment.

Defining code deployment

Code deployment can depend on the scale of your organization. Maryann and Jackie both work at large companies (Adobe and Spotify, respectively), where deployment likely differs from an individual developer's or a small organization's processes.

Jackie says, "At a high level, deploying code is taking the code that you've written and putting it into all of these places. But in practice, for a product and an organization at the scale of Spotify, code deployment is taking it from one place and putting it into lots of places. Deploying code isn't necessarily the same as releasing code." 

Code deployment risks

The risks when deploying code are varied. You might accidentally bring down your app or website, or deploy a bug system-wide. According to Jackie, less obvious but more sinister risks include not knowing how to manage an incident if something happens.

Some developers feel hesitant about deploying code, and with good reason. Jackie recalls an incident where her team made last-minute changes that ended up taking an entire site down. The experience made her hesitant about deploying code for a while, but she says it's best to just get back on the horse. "The important part is that you learn from those experiences and put improvements into place so that you can build that confidence back up."

Deploy on Friday with confidence

Deploying code on Friday is often avoided because if issues arise on the weekend, no one is available to make fixes. Both Jackie and Maryann have tips to reduce this risk. 

• Rollbacks allow you to go back to the last safe state with the press of a button

Deploy your technology on Friday on Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

What is Jetty?

Today, our guests share their extensive experience with Jetty. Greg Wilkins is the original software engineer for Jetty and Simone Bordet is a Jetty Committer. Jetty itself is an HTTP server and container for deploying Java servlets that run on HTTP, and it was also the first Java application server to be deployed as a clickable JAR file. Jetty started as a small open source project, and then moved to the Eclipse Foundation in 2009, where it still finds its home today.

Greg adds, “Jetty is actually a lot more than an HTTP server. Over the years, the way we've developed the software is rather than being an application container, Jetty is a good software component first that can be used to make an application server, but it can also be used to make embedded applications and various other things.” 

Who uses Jetty?

Jetty has a 25-year history with clients big and small, commercial and open source, from SaaS products to PaaS services. As Simone puts it, “Jetty is deeply battle-tested.” What do clients like about Jetty? It is:

• Highly extensible, flexible, and pluggable

What’s ahead for Jetty?

The new Jetty comes with significant improvements in performance and documentation, new features, and upgrades to Java 11 and Jakarta EE 9. Moving forward, Simone and Greg have plenty more they’d like to explore. “In the future, I'd like to see the standard support a good, reactive and/or asynchronous, scalable HTTP protocol and all the protocol features, and then have the application features layered on top of that,” says Greg. 

After working on Jetty for so long, Greg says it is just as relevant as it ever was. “It is amazing to keep doing it and to see the way Jetty keeps developing. What happens above us and the way people use the web changes all the time, but they still need this HTTP protocol.” 

Try out Jetty on Platform.sh: https://github.com/platformsh-templat... 

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

“97 Things Every Cloud Engineer Should Know” is a book of collected wisdom by cloud engineering experts. Our two guests today, Michelle Brenner and Dan Moore, both contributed chapters, and we speak with them about the book and the state of cloud engineering in 2021.

What is cloud engineering?

With most companies having at least part of their infrastructure in the cloud, some form of cloud engineering knowledge is necessary. Dan says, “I would say that a cloud engineer is someone who works in the cloud — public or private. Most people are cloud engineers nowadays, whether they want to be or not.”

Michelle expands on this definition. “A cloud engineer is anyone who wants their application not on their computer and more widely available. Whether it's at a company internally or externally for the whole world to try, it’s just kind of getting it out there and being more widely accessible.”

Gain an edge with managed services 

Dan and Michelle are big proponents of managed services, that is outsourcing tasks to people who know how to solve a given problem better than you do. They acknowledge that it’s a trade-off, but the reduced time-to-market and mental load can make it worthwhile.

Dan says, “It's just so exciting to me, as a developer, to be able to let something that was previously a highly specialized job be taken care of by these specialists. I just don't have to worry about certain aspects and I can focus on building things that I only I could build.”

Become a better engineer

Dan and Michelle come from different backgrounds and levels of experience, but they both have great advice for engineers. Try to have a:

• 15-minute rule (or 30) — Both Dan and Michelle have a time limit when it comes to solving problems, and after it expires, seek help from someone else.

Get your copy of 97 Things Every Cloud Engineer Should Know to learn more about modern cloud engineering from the experts.

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Our guests today Rudy de Busscher, on the Payara Sales team, and Fabio Andres Turizo, a Payara Engineer, speak with us about the importance of standards, what Jakarta EE offers developers, and using Payara. 

Defining Payara

Payara is a cloud-native, open source middleware platform that’s both Jakarta EE and MicroProfile compatible. It comes in two versions; community and enterprise. With the enterprise version, you get access to partners in the community, and very long-term support — 10 years! Payara supports on-premise, in the cloud, and hybrid Jakarta EE applications.

Standards mean interoperability and sustainability

Both Rudy and Fabio are big on standards, especially when it comes to microservices development. Fabio says, “Standards are important for multiple reasons — but I think the main one is variety. Where there's a body for standards, there's room for anyone to develop an implementation of that standard. And you as a developer have the option to choose what it is.”

There may be many reasons you can’t continue using a specific technology. In those cases, Fabio says, “Following a set of standards guarantees that you can quickly migrate to another vendor, and that migration is easier because both vendors are following the same standard. The process becomes more pain-free.”

Payara in the community

Payara is a successor to the now-defunct Glassfish. But Payara has some things Glassfish did not, according to our guests:

• Higher code quality

More comprehensive documentation lies ahead!

One of Payara's goals for 2021 is to make their documentation even more inclusive and welcoming. Fabio says, “One of the main plans for the year is to integrate everything — make it easily readable and more intuitive. If you're just starting out, or you’re a mid-level engineer trying to understand the nuances of how to operate Payara properly, then you will have all the tools you need in the documentation.”

Try Payara on Platform.sh: https://platform.sh/marketplace/templ...

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Today we’ll be talking about reactive programming, Quarkus and Mutiny with our experts, Clement Escoffier and Julien Ponge, both Principal Software Engineers at Red Hat.

Why use reactive programming

Reactive programming differs from the “traditional” imperative paradigm. Reactive is a programming approach that centers on events (and reacting to them!). It helps build robust, efficient, concurrent applications and systems, and it lets you handle more load while using resources more efficiently. As Clement points out, reactive also takes a different approach to failures. “Failures are inevitable. So we need to embrace them and be able to handle them gracefully.” He says.

So, why go reactive? Clement sums it up admirably. “With reactive, we are trying to build more responsive, efficient, robust, distributed systems. It’s about doing more with less.” 

Mutiny simplifies the development of reactive applications

Mutiny is a new reactive programming library built to bypass common issues with reactive programming. It is integrated — but not bound — into Quarkus, a  commonly used framework for building reactive applications. Mutiny helps developers by being:

• Event-driven

Julien adds that another of Mutiny’s strengths is that it’s built based on real-world scenarios. “We asked questions and got lots of feedback from real organizations on how Mutiny is going to be used.”

Clement sums it up. “With Mutiny, what we really wanted to tackle is a better user experience, an effective way to write non-blocking code, and make composing asynchronous operations easy and understandable.”

Get started with reactive programming

Clement and Julien both recommend the Get Started Guides for Quarkus and Mutiny. Julien says, “The guides help explain the concepts and contain repositories you can follow along with.” Clement adds a caveat; “Don’t write reactive just to write it; only do it if you have a need.”

But if you do need it, “For reactive programming, Quarkus and Mutiny are complete ecosystems that have everything you need.” Clement says.

Get started with Quarkus on Platform.sh.

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Business automation has been used in other industries for years.  Now it’s available for software development. Our Red Hat guests Karina Varela and Donato Marrazzo tell us how business automation can help bridge the gap between business and technical teams.

What is business automation?

According to Red Hat, “Business automation is the alignment of business process management (BPM) and business rules management (BRM) with modern application development to meet changing market demands.” 

Karina and Donato add their own definitions to the mix. Donato says, “Business automation is a bundle of two well-known technologies: one is the business process management (BPM) and the other one is digital management. When you contract these two, you are automating your business logic.” Karina adds, “When we are talking about automating business logic for decision processes, that’s business automation.”

The Kogito framework

Kogito, a cloud-native framework, is part of Red Hat’s business automation stack. “Kogito is this initiative that’s trying to modernize all our middleware, all of our processes, rules, and optimization, and make it even more lightweight, to make it run on top of a distributed environment instead of being in a monolith environment,” says Donato. Kogito appeals to developers for several reasons:

• Uses Quarkus to enable fast boot times and easier scaling 

Mathematical optimization with OptaPlanner

Karina and Donato tell us about a relative newcomer to the Red Hat business automation portfolio, OptaPlanner, which focuses on mathematical optimization. Some real-world use cases for OptaPlanner include:

• Assigning shifts at a busy hospital

As Donato says, “Finding the optimal solution is nearly impossible, but finding the near-optimal solution, that’s what OptaPlanner is for. It’s constraint solving with artificial intelligence.” Karina explains more on how OptaPlanner works and how to use it, “You have to design the model and the constraints, and the OptaPlanner engine is going to solve the problem for you. ”

Try Quarkus on Platform.sh

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Composer 2.0

Jordi Boggiano and Nils Aderman are the original authors of Composer, a near-universal package and dependency manager for PHP, and it just hit a major milestone: Composer 2.0. The two join us in this episode to give us the download on Composer 2 improvements and updates. 

But first, Nils helpfully defines Composer for us. “Composer is a package manager, or as we sometimes refer to it, a dependency manager. It’s responsible for installing and managing your project’s different versions of dependent packages.”

The Packagist-Composer relationship 

How does Packagist tie into Composer? Nils explains, “Composer is the command-line tool, or potentially the code inside it, that can be used as a library for integration into some web servers. And Packagist is the server, the registry, the repository for packages that you can download.”

Composer’s unique features and predictability 

Composer is not the first or the most recent language package manager on the market, but it does some things really well that other packages don’t.

• Handles automatically releasing packages with tags

In addition to the log file, our guests say predictability is a very deliberate component of Composer. Nils says, “Overall, the tools behaving in as predictable a way as possible is something that we agree on. I don't want to have to sit there and think, “Why the hell did it do this? I don't understand.” And then I have to read the docs for a couple of hours just to understand what is going on.”

Composer 2.0 improvements

The most common issues in Composer 1 were memory and performance issues. The performance issue has definitely been resolved in Composer 2; as Jordi says, “In Composer 2, the performance is so much faster and uses so much less memory. The numbers vary a bit, but there is a huge reduction of runtime and memory usage.” 

In addition to the performance overhaul, upgrading to Composer 2.0, also comes with:

• Internally refactored dependency updates and automatic installs

Try Packagist on Platform.sh

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

What is the Platform.sh Partner Program

For digital agencies, Platform.sh offers a number of incentives and benefits that make our service more attractive and your business more profitable. The team behind the Digital Agency Partner Program, and our guests today, explain how this partner program includes exclusive pricing, co-selling, and special features. 

Revamping the partner program

We brought on Chantal Pastorek to revamp and relaunch the Platform.sh partner program. Now, as our Agency Partner Manager for North America, she explains the value proposition we’re looking for, “We wanted to find a way to really identify the agency partners that were going to work with us to co-sell, and to achieve goals for both the agency and Platform.sh. We wanted to figure out who our actual partners were and clean that up across our systems.”

Identifying and expanding the partner profile

Chantal touches on the diversity of our partner organizations, “The types of agencies we want to work with are people building websites, managing multiple projects, or building out different web applications.” 

Aaron Porter, Partner Program Global VP, credits the visionary architecture of the Partner Program toolset for not limiting our audience. “We don't have to be a tool or a program that's focused on one particular type of agency. We have the capabilities to take on partners that are still growing in complexity or are already a large agency.” 

The Partner Portal: your one-stop partnership hub

Chantal gives the lowdown on what’s included in our partner portal. “The partner portal was created as a resource for partners, so they could access training, register and track their opportunities, but also access materials for co-selling, competitive intel, and marketing assets.” In the product dashboard, we share with partners what’s being released, what’s coming up, and opportunities to provide feedback, input, and ideas.

Sell more, sell better: partner enablement

For Aaron, “enablement” is the key word. Every step of our partner program is about enabling partners to sell more and better to their clients. “Whether it's technical expertise training, a strategic kind of selling session, a pre-sales session, or co-selling live, each one of those pieces of enablement is what gets their project over the goal line and what drives the partner program. We want to powerfully aid our partners in their success.” 

Try the Platform.sh Partner Program to enable your agency’s success

Platform.sh

Platform.sh on social media

Watch, listen, subscribe to the Platform.sh Deploy Friday podcast:

Platform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.