
Sign up to save your podcasts
Or


We previously released this episode with the wrong audio file and are re-releasing it on a weekend.
TIBCO was started in the 90’s with a popular message bus product that was widely used by finance companies, logistics providers, and other systems with high throughput. As TIBCO grew in popularity, the company expanded into other areas through products it developed in-house as well as through acquisitions.
One acquisition was Jaspersoft, a business intelligence data platform. When TIBCO acquired Jaspersoft in 2014, the architecture was a monolithic Java application. Around this time, customer use cases were shifting from centralized reporting to real-time, embedded visualizations.
The use case of the Jaspersoft software was becoming less centralized and less monolithic and the software architecture needed to change in order to reflect that.
Jan Schiffman is a VP of engineering at TIBCO and Sherman Wood is a director at TIBCO. They join the show to discuss the process of migrating a large Java monolith to a composable set of services. Breaking up a monolith is not an easy process–nor is it something that every company should do just because they have a monolith. In some cases, a monolith is just fine.
Jan and Sherman explain why the business use case for why the Jaspersoft monolith needed to be refactored, and their approach to the refactoring. We also talk through the modern use cases of embedded analytics and the interaction between business analysts and data engineers. At a higher level, we discuss the lessons they have learned from managing a large, complex refactoring. Full disclosure: TIBCO is a sponsor of Software Engineering Daily.
Linkedin is an organization with thousands of employees. An enterprise of that size starts to develop problems with data collaboration. Data collaboration is the process of sharing and analyzing data with multiple users, such as data scientists, business analysts, and engineers.
How do data scientists know what questions to ask? How do business analysts know the right way to query a database? How does a data engineer even find where the right database is within the company infrastructure? And how can these different users share information with each other so that redundant work is avoided?
When Adam Weinstein was at Linkedin, he saw these problems firsthand. The process of accessing and utilizing data felt slow and broken. Engineers were searching through a company wiki to find out how to leverage data, and the wiki was often out of date. When an engineer would leave the company, there was not a durable, institutional memory of how that engineer worked with data.
Adam used this experience as inspiration for Cursor, a tool for data collaboration. Cursor allows different users in the data pipeline to share data sets, queries, access patterns, and comments about data within a company. Cursor is used by Linkedin, Slack, Apple, and other companies. Adam is the CEO of Cursor, and he joins the show for an interview about the problems and opportunities of data collaboration.
Internet of Things is a concept that describes lots of devices that you interact with regularly being connected to the Internet and networked together.
Technologists have been dreaming of the world of IoT for many years, where our connected refrigerator can detect that we are out of food, and automatically order more food. Or our connected bathroom can scan us for diseases and recommend treatment.
The bright future of IoT is slowly coming together. Hardware prototyping is getting cheaper. Voice interfaces and machine learning are creating new mediums for communicating with devices. Platforms like Kickstarter are allowing developers to validate the market for their products and raise the necessary capital to build their product.
Android Things is a developer platform for IoT applications based on the Android Operating system. Android Things consists of hardware devices and software tools that reduce common IoT problems such as software updates and security patches. Wayne Piekarski is a staff developer advocate at Google, and he joins the show to talk about the state of IoT and why Google built Android Things.
JavaScript performance has improved over time due to advances in JavaScript engines such as Google’s V8. A JavaScript engine performs compiler optimization, garbage collection, hot code management, caching, and other runtime aspects that keep a JavaScript program running efficiently. JavaScript runs in browsers and servers. The resources that are available to a JavaScript engine vary widely across different machines.
JavaScript code is parsed into an abstract tree before being handed off to the compiler toolchain, in which one or more optimizing compilers produce efficient low-level code. In recent shows about WebAssembly, we have covered compiler pipelines. In an episode about GraalVM, we explored the impact that “code shape” has on the efficiency of JavaScript execution.
Mathias Bynens is a developer advocate at Google working on the V8 JavaScript engine team. In today’s show we explore how a JavaScript engine works, and how compiler toolchains can adapt the hot code paths depending on what code needs to be optimized for.
Front is a shared inbox application that has seen rapid adoption within companies. Front allows multiple members of a company to collaborate together on a conversation–whether that conversation is in email, Twitter, or Facebook Messenger. This is useful when a customer email needs to be shared between the sales and engineering teams, or when a single email address is shared between different members of the same team, such as “[email protected]”.
This might sound like a niche problem, but it is actually a problem faced somewhere within every single company. Because the problem of shared inbox is so prevalent, the company has grown its user base quickly, scaling the team as well as the infrastructure.
The sensitivity of the data (emails) that Front is handling means that security is paramount. And as users of Front rely on it more and more as a central point of communication, uptime and consistency needs to be maintained.
Laurent Perrin is the CTO at Front, and he joins the show to describe the software architecture and product strategy for Front. It was a fascinating show, and we covered the full stack. On the backend, Front pulls emails into S3 buckets and maintains the schema of the inbox in a SQL database. The desktop Front client is written in Electron, which is a way to write desktop applications in HTML5, JavaScript, and CSS.
We also talked about the system for keeping the communications “real-time”–it’s important that users are aware of what each other is doing, since you don’t want to be preparing a response to an email at the same time I am.
Engineering organizations can operate more efficiently by working with a continuous integration and continuous deployment workflow. Continuous integration is the process of automatically building and deploying code that gets pushed to a remote repository. Continuous deployment is the process of moving that code through a pipeline of environments, from dev to test to production. At each stage, the engineers feel increasingly safe that the code will not break the user experience.
When a company adopts Kubernetes, the workflow for deploying software within that company might need to be refactored. If the company starts to deploy containers in production, and managing those containers using Kubernetes, the company will also want to have a testing pipeline that emulates the production environment using containers and Kubernetes.
Sheroy Marker is the head of technology at ThoughtWorks products, where he works on GoCD, a continuous delivery tool. Sheroy joins the show to talk about how Kubernetes affects continuous delivery workflows, and the process of building out Kubernetes integrations for GoCD.
We also discussed the landscape of continuous delivery tools–why there are so many continuous delivery tools, and the question of how to choose a continuous delivery product if you are implementing CD. Continuous delivery tooling is in some ways like the space of monitoring, logging, and analytics–there are lots of successful products in the market. Full disclosure: ThoughtWorks and GoCD are sponsors of Software Engineering Daily.
Engineering organizations can operate more efficiently by working with a continuous integration and continuous deployment workflow. Continuous integration is the process of automatically building and deploying code that gets pushed to a remote repository. Continuous deployment is the process of moving that code through a pipeline of environments, from dev to test to production. At each stage, the engineers feel increasingly safe that the code will not break the user experience.
When a company adopts Kubernetes, the workflow for deploying software within that company might need to be refactored. If the company starts to deploy containers in production, and managing those containers using Kubernetes, the company will also want to have a testing pipeline that emulates the production environment using containers and Kubernetes.
Sheroy Marker is the head of technology at ThoughtWorks products, where he works on GoCD, a continuous delivery tool. Sheroy joins the show to talk about how Kubernetes affects continuous delivery workflows, and the process of building out Kubernetes integrations for GoCD.
We also discussed the landscape of continuous delivery tools–why there are so many continuous delivery tools, and the question of how to choose a continuous delivery product if you are implementing CD. Continuous delivery tooling is in some ways like the space of monitoring, logging, and analytics–there are lots of successful products in the market. Full disclosure: ThoughtWorks and GoCD are sponsors of Software Engineering Daily.
TIBCO was started in the 90’s with a popular message bus product that was widely used by finance companies, logistics providers, and other systems with high throughput. As TIBCO grew in popularity, the company expanded into other areas through products it developed in-house as well as through acquisitions.
One acquisition was Jaspersoft, a business intelligence data platform. When TIBCO acquired Jaspersoft in 2014, the architecture was a monolithic Java application. Around this time, customer use cases were shifting from centralized reporting to real-time, embedded visualizations.
The use case of the Jaspersoft software was becoming less centralized and less monolithic and the software architecture needed to change in order to reflect that.
Jan Schiffman is a VP of engineering at TIBCO and Sherman Wood is a director at TIBCO. They join the show to discuss the process of migrating a large Java monolith to a composable set of services. Breaking up a monolith is not an easy process–nor is it something that every company should do just because they have a monolith. In some cases, a monolith is just fine.
Jan and Sherman explain why the business use case for why the Jaspersoft monolith needed to be refactored, and their approach to the refactoring. We also talk through the modern use cases of embedded analytics and the interaction between business analysts and data engineers. At a higher level, we discuss the lessons they have learned from managing a large, complex refactoring. Full disclosure: TIBCO is a sponsor of Software Engineering Daily.
Engineering organizations can operate more efficiently by working with a continuous integration and continuous deployment workflow. Continuous integration is the process of automatically building and deploying code that gets pushed to a remote repository. Continuous deployment is the process of moving that code through a pipeline of environments, from dev to test to production. At each stage, the engineers feel increasingly safe that the code will not break the user experience.
When a company adopts Kubernetes, the workflow for deploying software within that company might need to be refactored. If the company starts to deploy containers in production, and managing those containers using Kubernetes, the company will also want to have a testing pipeline that emulates the production environment using containers and Kubernetes.
Sheroy Marker is the head of technology at ThoughtWorks products, where he works on GoCD, a continuous delivery tool. Sheroy joins the show to talk about how Kubernetes affects continuous delivery workflows, and the process of building out Kubernetes integrations for GoCD.
We also discussed the landscape of continuous delivery tools–why there are so many continuous delivery tools, and the question of how to choose a continuous delivery product if you are implementing CD. Continuous delivery tooling is in some ways like the space of monitoring, logging, and analytics–there are lots of successful products in the market. Full disclosure: ThoughtWorks and GoCD are sponsors of Software Engineering Daily.
Stitch Fix is a company that recommends packages of clothing based on a set of preferences that the user defines and updates over time. Stitch Fix’s software platform includes the website, data engineering infrastructure, and warehouse software. Stitch Fix has over 5000 employees, including a large team of engineers.
Cathy Polinsky is the CTO of Stitch Fix. In today’s show Cathy describes how the infrastructure has changed as the company has grown–including the process of moving the platform from Heroku to AWS, and the experience of scaling and refactoring a large monolithic database. Cathy also talked about the management structure, the hiring process, and engineering compensation at Stitch Fix.
From the publisher's feed