NewsTech News

Software Engineering Daily

softwareengineeringdaily.com

Software Daily: The world through the lens of software.

Play

Open in

2,221 episodes
Avg 53 min
English

Episodes 2,221

Newest Oldest Longest

All20 Unplayed20 In progress0 Played0

Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.

#882
Aug 27, 2018 · 52 min
Helm with Michelle Noorali
Back in 2014, platform-as-a-service was becoming an increasingly popular idea. The idea of PaaS was to sit on top of infrastructure-as-a-service providers like Azure, AWS, or Google Cloud, and simplify some of the complexity of these infrastructure providers. Heroku had built a successful businesses from the idea of platform-as-a-service, and there was a widely held desire in the developer community to have an “open source Heroku.” One project that was working towards the idea of an open source platform-as-a-service was Deis. Deis made it easier for people to deploy and manage their applications, and it simplified some of the hard parts of container management. When Kubernetes came out, Deis got refactored to use Kubernetes under the hood for container orchestration. Deis was one of the first projects to use Kubernetes as a tool to build a platform-as-a-service, and the team that was working on Deis got very early exposure to the process of building a platform on top of Kubernetes. Michelle Noorali was one of the engineers on the Deis team. When Deis got acquired by Microsoft, Michelle was working on Helm, a package manager for distributed systems. Helm allows developers to deploy distributed applications on top of Kubernetes more easily. A few examples of distributed applications that can be deployed using Helm are Kafka, Prometheus, and IPFS. One reason Helm is so useful is that distributed systems are notoriously hard to configure and run. Since joining Microsoft, Michelle has continued to work on Helm. She is also a member of the Kubernetes Steering Committee and the board of the CNCF. Michelle joins the show to talk about her early experiences building PaaS and her perspective on the Kubernetes ecosystem. Full disclosure: Microsoft is a sponsor of Software Engineering Daily.
#881
Aug 23, 2018 · 53 min
StitchFix Engineering with Cathy Polinsky
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.
#880
Aug 21, 2018 · 42 min
OLIO: Food Sharing with Lloyd Watkin
Food gets thrown away from restaurants, homes, catering companies, and any other place with a kitchen. Most of this food gets thrown away when it is still edible, and could provide nutrition to someone who is hungry. Just like Airbnb makes use of excess living capacity, OLIO was started to connect excess food with people who want to eat that food. There are numerous challenges with this idea. How do you control quality and ensure the food is safe? How do you make money as a business? How do you solve the chicken and egg problem, and make sure that you get hungry users and people with food to give away at the same time? Lloyd Watkin is a software engineer at OLIO, and he joins today’s episode to describe how the platform works, how it is built, and how the company plans to scale their large base of volunteers. It’s a fascinating set of operational and engineering issues.
#879
Aug 21, 2018 · 1 hr 3 min
Build Faster with Nader Dabit
Building software today is much faster than it was just a few years ago. The tools are higher level, and abstract away tasks that would have required months of development. Much of a developer’s time used to be spent optimizing databases, load balancers, and queueing systems in order to be able to handle the load created by thousands of users. Today, scalability is built into much of our infrastructure by default. We have had several years of infrastructure with automatic scalability, and some of the more recent advances in developer tooling are about convenience, and faster development time. Developers are spending less time dealing with the ambiguous idea of a “server” and more time interacting with well-defined APIs and data sources. A few examples are AppSync from Amazon Web Services and Firebase from Google. These tools are like databases with rich interactive functionality. Instead of having to create a server to listen to a database for changes and push notifications to users in response to those changes, AppSync and Firebase can be programmed to have this kind of functionality built in. There are many other examples of high level APIs, rich backends, and developer productivity tools that lead to shorter development time. What does this mean for developers? It means we can build much faster. We can prototype quickly for low amounts of money–without sacrificing quality. We can spend more time focusing on design, user experience, and business models and less time focusing on keeping the application up and running. Nader Dabit is a developer advocate at Amazon Web Services, and he returns to the show to discuss modern tooling, and how that tooling changes the potential for high output and fast iteration among developers. It is a strategic, philosophical discussion of how to build modern software.
#878
Aug 20, 2018 · 58 min
WebAssembly Engineering with Ben Smith and Thomas Nattestad
WebAssembly allows developers to run any language in a sandboxed, memory controlled module that can be called via well-defined semantics. As we have discussed in recent episodes with Lin Clark and Steve Klabnik from Mozilla, WebAssembly is changing application architectures both in and outside the browser. WebAssembly is being adopted by all of the major browser vendors, including Google. Today’s guests are Thomas Nattestad and Ben Smith from Google. Thomas is the PM for V8, WebAssembly, Storage, and Games on the web and Ben is a software engineer on the Chrome team. Ben and Thomas talk about the state of WebAssembly, what the different browser manufacturers are doing, and some cool uses for WebAssembly–from games to CDNs to cryptocurrency infrastructure. As in the previous episodes, there was discussion of WebAssembly’s security and memory benefits from being a bounded context.
#877
Aug 17, 2018 · 56 min
WebAssembly Future with Steve Klabnik
WebAssembly is a low-level compilation target for any programming language that can be interpreted into WebAssembly. Alternatively, WebAssembly is a way to run languages other than JavaScript in the browser. Or, yet another way of describing WebAssembly is a virtual machine for executing code in a low level, well-defined sandbox. WebAssembly is reshaping what is possible to do in the web browser. A developer can write a program in Rust or C++, compile it down into a WebAssembly module, and call out to that module via JavaScript. This is very useful for running memory-sensitive workloads in the browser—such as 3-D games. But WebAssembly is also useful for running modules outside of the browser. Why is that important? If you can already run C++ or Rust code outside of the browser by executing the program from the command line, why would you want to put the code into a WebAssembly module before executing it? One answer is security. WebAssembly modules have well-defined semantics for what memory they can access. WebAssembly could provide more reliable sandboxing for untrusted code. Steve Klabnik is a software engineer at Mozilla, and he joins the show today to play the role of a WebAssembly futurist. We revisit the basics of WebAssembly and the current state of the technology. Steve talks Steve also describes the lessons of past web technologies such as Flash—and what they did right and wrong. We also explore the current and future applications of WebAssembly, which we will talk about in more detail in tomorrow’s episode.
#876
Aug 16, 2018 · 54 min
DoorDash Engineering with Raghav Ramesh
DoorDash is a logistics company that connects customers, restaurants, and drivers that can move food to its destination. When a customer orders from a restaurant, DoorDash needs to identify the ideal driver for picking up the order from the restaurant and dropping it off with the customer. This process of matching an order to a driver takes in many different factors. Let’s say I order spaghetti from an Italian restaurant. How long does the spaghetti take to prepare? How much traffic is there in different areas of the city? Who are the different drivers who could potentially pick the spaghetti up? Are there other orders near the Italian restaurant, that we could co-schedule the spaghetti delivery with? In order to perform this matching of drivers and orders, DoorDash builds machine learning models that take into account historical data. In today’s episode, Raghav Ramesh explains how DoorDash’s data platform works, and how that data is used to build machine learning models. We also explore the machine learning model release process—which involves backtesting, shadowing, and gradual rollout.
#875
Aug 15, 2018 · 1 hr 1 min
Casa: Crypto Wallet Security with Jameson Lopp
Cryptocurrency security is a concern to anyone who has a significant amount of money in the form of Bitcoin, Ethereum, or other crypto assets. Most Bitcoin is held in either a Bitcoin wallet or a Bitcoin bank. Your Bitcoin holdings are recorded on a public ledger. You access these holdings by authenticating with your private key. A Bitcoin wallet could be described more accurately as a Bitcoin keyring. Securing your Bitcoin wallet is about securing that private key. Just as there are many different ways to secure any individual piece of text, there are many ways to secure a Bitcoin private key. A Bitcoin “bank” is a term that can be used to describe institutions such as Coinbase. Coinbase takes the technology of the Bitcoin wallet and wraps it in additional layers of security, identity, and failover that we associate with banks and large technology companies. By using a Bitcoin bank, you sacrifice the autonomy of managing your own private key. On the bright side, you don’t have to manage your own private key. If you lose your Coinbase password, there are plenty of ways to recover it. A Bitcoin bank gives you the downsides and the upsides of working with a centralized service provider. Jameson Lopp is a cypherpunk and cryptocurrency engineer at Casa. Casa is a company that is building long-term cryptocurrency storage and secure key infrastructure. In this episode, we explore how Bitcoin wallets work, how to secure them, the common threats, scams and hacking attempts of Bitcoin, and what he is working on at Casa.
#874
Aug 14, 2018 · 51 min
Infrastructure Monitoring with Mark Carter
At Google, the job of a site reliability engineer involves building tools to automate infrastructure operations. If a server crashes, there is automation in place to create a new server. If a service starts to receive a high load of traffic, there is automation in place to scale up the instances of that service. In order to create an automated response to an infrastructure problem, a site reliability engineer needs insights into that infrastructure. Every service needs tools around monitoring, alerting, debugging, and distributed tracing. One benefit of working at a large company like Google is that an engineer building a new product gets this kind of tooling by default. If I am hacking on a project at home, I have to set up all kinds of tools to help me diagnose and resolve problems. Setting up this tooling takes time, and requires expertise. Stackdriver is a set of tools and instrumentation that allows developers to monitor, debug, and inspect infrastructure. Stackdriver is based on the internal observability tools built for Google. Mark Carter is a group product manager at Google, and he joins the show to discuss site reliability engineering and the creation of Stackdriver.
#873
Aug 13, 2018 · 44 min
GitOps: Kubernetes Continuous Delivery with Alexis Richardson
Continuous delivery is a way of releasing software without requiring software engineers to synchronize during a release. Over the last decade, continuous delivery workflows have evolved as the tools have changed. Jenkins was one of the first continuous delivery tools and is still in heavy use today. Netflix’s open sourced Spinnaker has also been widely adopted. As Kubernetes has grown in popularity, some engineers have developed a workflow around Kubernetes and Git known as GitOps. GitOps treats Git as the source of truth for deployments. Under GitOps, when a divergence occurs between your git repository’s configuration files and the state of your production infrastructure, your infrastructure should automatically adjust its state to align with the state defined in git. Alexis Richardson is the CEO of Weaveworks, a company that has built tooling around GitOps. He joins the show to describe how GitOps works, and explain how it compares to other methods for continuous delivery.
#872
Aug 10, 2018 · 47 min
Klarna Engineering with Marcus Granström
Klarna is a payments company headquartered in Sweden. Since being established in 2005 it has grown to handling $21 billion in online sales in 2017. Roughly 40% of all e-commerce sales in Sweden go through Klarna. Klarna’s original differentiator was that it allowed users to checkout of e-commerce stores without entering in credit card information. Instead, the user enters an email address and registers with Klarna. This allows Klarna to assume the risk of the transaction, in place of the credit card company. Klarna’s clever payment method became very popular, and 13 years later Klarna is a bank with a variety of financial services and payment methods. Marcus Granstrom is a director of engineering at Klarna. His work ranges from product development to systems architecture to management. His cross functional role has some similarity to Raylene Yung from Stripe, who is also an engineering director at a payments company, and was on the show yesterday. Marcus walked me through the life of a payment hitting Klarna’s servers, and this served as a nice starting point for a conversation about Klarna’s infrastructure, their product, and their engineering practices.
#871
Aug 9, 2018 · 44 min
Stripe Engineering with Raylene Yung
Stripe is a payments API that allows merchants to transact online. Since the creation of the payments API, Stripe has expanded into adjacent services such as fraud detection, business management, and billing. These other verticals leverage the existing customer base and infrastructure that Stripe has developed from the success of their payments business. Raylene Yung is the head of payments at Stripe. She joins the show to talk about her work, which includes elements of engineering, product development, design, and management. All of these dimensions of her job came up in our conversation, which made for a wide ranging conversation. This interview comes in the context of Stripe’s rapid growth. The organization is changing, and Raylene explored the questions that Stripe is asking itself internally about org structure. Namely: what is the tradeoff between a defined, hierarchical structure of direct reports versus a decentralized, flat org structure? Is there any advantage to making roles highly defined (such as “senior infrastructure software engineer”)? Or is it better to let people have fluid roles, and self-assemble? Raylene was willing to explore these questions–and I found her answers highly useful and thought provoking.
#870
Aug 8, 2018 · 59 min
Self-Driving Engineering with George Hotz
In the smartphone market there are two dominant operating systems: one closed source (iPhone) and one open source (Android). The market for self-driving cars could play out the same way, with a company like Tesla becoming the closed source iPhone of cars, and a company like Comma.ai developing the open source Android of self-driving cars. George Hotz is the CEO of Comma.ai. Comma makes hardware devices that allow users with “normal” cars to be augmented with advanced cruise control and lane assist features. This means you can take your own car–for example, a Toyota Prius–and outfit your car to have something similar to the Tesla Autopilot. Comma’s hardware devices cost under $1000 to order online. George joins the show to explain how the Comma hardware and software stack works in detail–from the low level interface with a car’s CAN bus to the high level machine learning infrastructure. Users who purchase the Comma.ai hardware drive around with a camera facing the front of their windshield. This video is used to orient the state of the car in space. The video from that camera also gets saved and uploaded to Comma’s servers. Comma can use this video together with labeled events from the user’s driving experience to crowdsource their model for self-driving. For example, if a user is driving down a long stretch of highway, and they turn on the Comma.ai driving assistance, the car will start driving itself and the video capture will begin. If the car begins to swerve into another lane, the user will take over for the car and the Comma system will disengage. This “disengagement” event gets labeled as such, and when that data makes it back to Comma’s servers, Comma can use the data to update their models. George is very good at explaining complex engineering topics, and is also quite entertaining and open to discussing the technology as well as other competitors in the autonomous car space. I have not been able to get many other people on the show to talk about autonomous cars, so this was quite refreshing! I hope to do more in the future.
#869
Aug 7, 2018 · 1 hr 4 min
Future Architecture with Chad Fowler
Chad Fowler was the CTO of Wunderlist prior to its acquisition by Microsoft. Since the acquisition, Chad has become the general manager of developer advocacy at Microsoft. He also works as a venture capitalist at BlueYard Capital, an early stage investment firm. I’ve had a lot of fun talking to Chad, because he can move seamlessly between talking about disparate subjects like cloud computing, investing, cryptocurrencies, and music composition. And he has novel things to say about all of them! When Chad joined Wunderlist, he helped start a large refactoring of the software architecture. He then helped the company navigate to the successful Microsoft acquisition. We started off the conversation with the story of this rearchitecture, and how he sees the current opportunities in front of Microsoft. Chad gives his perspective on Kubernetes, functions-as-a-service, and how developer tooling might evolve in the near future. After talking about near-term developer tooling, we talked about the distant future: bug bounty marketplaces on the blockchain; using Github repositories to train machine learning models about how to write code; the comparison between music collaboration and software collaboration. This was a wide array of topics, but Chad was equipped to discuss all of them–since he works at Microsoft, makes large investments in the future, and studied music when he was in school.
#868
Aug 6, 2018 · 59 min
Splice: Music Collaboration with Matt Aimonetti
Music collaboration has historically been accomplished by musicians gathering in bands. A band is usually an in-person, physical manifestation: a drummer, a guitarist, a piano player. Or, on a large scale, a symphony of classical instruments led by a conductor. Today, the most flexible instrument that anyone can play is arguably the computer, because a computer can simulate or replay any of the sounds made by any other instrument. Another advantage of the computer is that it removes physicality as a constraint on the musician. A computer musician does not have to train their muscles to play piano, or guitar, or drums. The computer musician can imagine a sound and bring it to life inside a digital audio workstation (a program for composing and arranging music). The rise of the computer musician has coincided with a change in the way popular music is created. Instead of bands needing to work together to create a piece of music, a single producer can simulate all of the members of the band by programming piano, drums, and everything else. The rise of the solo producer has given birth to new kinds of music–but solo music production inherently limits the range of musical ideas that can be explored. The most important works of art have input from multiple people. And even the most successful solo producers love to work with other artists who have a complementary skill–such as vocals. For the last twenty years, the model of solo producer working with pop vocalist has largely dominated the charts. Musical collaboration has stuck to a model that mimics its pre-Internet form, with very small groups of 1-5 people making the core of a song. The main tools that people use to collaborate are email and Dropbox. Splice is a tool for music collaboration. Splice combines version control, revision history, social networking, sample discovery, synthesizer rental, and other features. Splice is changing the way that music is created, with a large percentage of top producers adopting it. The impact Splice has on music will be on par with what Github has done for software engineering. Matt Aimonetti is the CTO and co-founder of Splice, and he joins the show to talk about the founding story, the product development, and the engineering of Splice.
#867
Aug 3, 2018 · 50 min
GraalVM with Thomas Wuerthinger
Java programs compile into Java bytecode. Java bytecode executes in the Java Virtual Machine, a runtime environment that compiles that bytecode further into machine code, and optimizes the runtime by identifying “hot” code paths and keeping those hot code paths executing quickly. The Java Virtual Machine is a popular platform for building languages on top of. Languages like Scala and Clojure compile down to Java bytecode, and can take advantage of the garbage collection system and the code path optimizations of the JVM. But when Scala and Clojure compile into Java bytecode, the code “shape”–the way that the programs are laid out in memory–is not the same as when Java programs compile into Java bytecode. Executing bytecode that comes from Scala will have certain performance penalties relative to a functionally identical program written in Java. GraalVM is a system for interpreting languages into Java bytecode that can run efficiently on the JVM. Any language can be interpreted into an abstract syntax tree that the GraalVM can execute using the JVM. Languages that can run on GraalVM include JavaScript, R, Ruby, and Python. Thomas Wuerthinger is a senior research director at Oracle and the project lead for GraalVM. He joins the show to explain the motivation for GraalVM, the architecture of the project, and the future of language interoperability. It was an exciting discussion and I learned a lot about the Java ecosystem.
#866
Aug 2, 2018 · 52 min
Token Types with Felipe Pereira
A token is a unit of virtual currency. Most tokens are built on a blockchain-based cryptocurrency platform, such as Ethereum. Building on top of a platform like Ethereum allows these tokens to form their own financial ecosystem while leveraging the scale of an existing currency. Tokens became highly popular in early 2018, with the boom in ICOs–initial coin offerings. Many of these coins offer a value proposition of a “utility token.” The idea of a utility token is that the token is necessary to transact in a particular ecosystem. If Amazon were to require you to convert US dollars to Amazon coins in order to buy items on Amazon, the Amazon coin would be a “utility token.” There are many different kinds of utility token schemes, and time will tell if this model makes sense for the cryptocurrency investment landscape. Another type of token is the “security token,” in which a token represents a share in an organization. This token type is more like a stock, or bond, or certificate of ownership of a financial instrument. These types of tokens also have their share of criticism. If I start a company, most of my assets are not represented on a blockchain–the assets are things like hiring contracts, intellectual property, real estate, etc. The legal ownership of these assets is settled by a complicated legal system which has no notion of a blockchain. It’s unclear how the claims of a security token today would be enforced–or why a security token is presently a better option for raising capital than traditional equity or debt instruments. Felipe Pereira is the author of “On the immaturity of tokenized value capture mechanisms,” a Medium article in which he documents different types of token systems, including several flavors of utility tokens and security tokens. He’s also the co-founder at a company called Paratii. He joins the show to discuss the present viability of token-based systems–and what blockchains have actually proven to be useful for today.
#865
Aug 1, 2018 · 52 min
Castor EDC with Derk Arts
Medical breakthroughs require medical research. Medical research requires patient testing and data collection. The most common form of capturing patient data is through surveys–and most of those surveys today are done on paper. Surveying patients to understand the side effects or benefits of trial drugs or treatments, and getting accurate results out of these are critical aspects of medical research. Traditionally, these surveys are filled and read manually, and entered into a database by a human operator. In these steps, there is too much room for human error, from unreadable handwritings to typos being entered into the databas Electronic Data Capture platforms were created out of this need for easy and accurate data collection for researches. By enabling online survey creation and result collection, EDC platforms improved medical research immensely. However, these platforms are complex to design. Where patient medical data is concerned, privacy and security are of extremely high importance. Compliance with laws that protect anonymity and privacy of the patients is necessary. On top of these, the platform must be easy-to-use, and reliable. Castor EDC is a company specializing on EDC for medical research, founded in the Netherlands and active in many countries around the globe. Our guest today is Derk Arts, the founder and CEO of Castor EDC. In this episode we discuss Electronic Data Capture platforms, how Castor EDC overcame the engineering and design problems, how they comply with the laws, and their business model.
#864
Jul 31, 2018 · 43 min
Jailbreaking Apple Watch with Max Bazaliy
Apple operating systems are closed source. This closed source nature gives Apple an extremely successful business model–and a very different software developer ecosystem than Linux-based systems. Since Linux is open source, the information on how to manipulate the system at a low level is very public. The lack of information about low-level programming in Apple operating systems has led to a large community of “jailbreaking”–where people try to reverse engineer how the closed source systems function. In today’s episode, Max Bazaliy joins the show to describe how he reverse engineered an Apple Watch. It’s a complex security challenge to jailbreak an Apple Watch, as he describes in detail. Max is a security researcher at Lookout, a mobile security company.
#863
Jul 30, 2018 · 53 min
Edge Kubernetes with Venkat Yalla
“Edge computing” is a term used to define computation that takes place in an environment outside of a data center. Edge computing is a broad term. Your smartphone is an edge device. A self-driving car is an edge device. A security camera with a computer chip is an edge device. These “edge devices” have existed for a long time now, but the term “edge computing” has only started being used more recently. Why is that? It is mostly because the volume of data produced by edge devices, and the type of computation that we want from edge devices is changing. We want to develop large sensor networks to enable smart factories, and smart agriculture fields. We want our smartphones to have machine learning models that get updated as frequently as possible. We want to use self-driving cars, and drones, and smart refrigerators to develop elaborate mesh networks–and perhaps even have micropayments between machines, so that computation can be offloaded from edge devices to a nearby mesh network for a small price. Kubernetes is a tool for orchestrating distributed, containerized computation. Just as Kubernetes is being widely used for data center infrastructure, it can also be used to orchestrate computation among nodes on-premise at a factory, or in a smart agriculture environment. In today’s episode, Venkat Yalla from Microsoft joins the show to talk about Kubernetes at the edge, and how Internet of things applications can use Kubernetes for their deployments today–and what the future might hold. Full disclosure: Microsoft is a sponsor of SE Daily.

Showing 1,401–1,420 of 2,221 episodes