
Sign up to save your podcasts
Or


Kubernetes has become the standard system for deploying and managing clusters of containers. But the vision of the project goes beyond managing containers. The long-term goal is to democratize the ability to build distributed systems.
Brendan Burns is a co-founder of the Kubernetes project. He recently announced an open source project called Metaparticle, a standard library for cloud-native development:
Metaparticle builds on top of Kubernetes primitives to make distributed synchronization easier… It supplies language independent modules for locking and leader election as easy-to-use abstractions in familiar programming languages.
After decades of distributed systems research and application, patterns have emerged about how we build these systems. We need a way to lock a variable, so that two nodes will not be able to write to that variable in a nondeterministic fashion. We need a way to do master election, so that if the master node dies, the other nodes can pick a new node to orchestrate the system.
We know that just about every distributed application needs locking and leader election–so how can we build these features directly into our programming tools, rather than bolting them on?
With Kubernetes providing a standard operating system for distributed applications, we can start to build standard libraries that assume we have access to underlying Kubernetes primitives. Instead of calling out to external tools like Zookeeper and etcd, a standard library like Metaparticle will abstract them away.
An example: if I am writing a system to do distributed mapreduce, I would like to avoid thinking about node failures and race conditions. Brendan’s idea is to push those problems down into a standard library–so the next developer who comes along with a new idea for a multi-node application has an easier time.
Brendan Burns currently works as a distinguished engineer at Microsoft, and he joins the show to discuss why it is still hard to build distributed systems and what can be done to make it easier. This is the second time we have had Brendan on the show. The first time he came on, he discussed the history of Kubernetes, and some of the design decisions of the system. This episode was more about the future. Full disclosure: Microsoft (where Brendan is employed) is a sponsor of Software Engineering Daily.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
You are requesting a car from a ridesharing service such as Lyft.
Your request hits the Lyft servers and begins trying to get you a car. It takes your geolocation, and passes the geolocation to a service that finds cars that are nearby, and puts all those cars into a list. The list of nearby cars is sent to another service, which sorts the list of cars by how close they are to you, and how high their star rating is. Finally, your car is selected, and sent back to your phone in a response from the server.
In a “microservices” environment, multiple services often work together to accomplish a user task. In the example I just gave, one service took a geolocation and turned it into a list, another service took a list and sorted it, and another service sent the actual response back to the user.
This is a common pattern: Service A calls service B, which calls service C, and so on. When one of those services fails along the way, how do you identify which one it was? When one of those services fails to deliver a response quickly, how do you know where your latency is coming from?
The solution is distributed tracing. To implement distributed tracing, each user level request gets a request identifier associated with it. When service A calls service B, it also hands off the unique request ID, so that the overall request can be traced as it passes through the distributed system (and if that doesn’t make sense–don’t worry, we explain it again during the show).
Ben Sigelman began working on distributed tracing when he was at Google and authored the “Dapper” paper. Dapper was implemented at Google to help debug some of the distributed systems problems faced by the engineers who work on Google infrastructure.
A request that moves through several different services spends time processing on each of those services. A distributed tracing system measures the time spent in each of those services–that time spent is called a span. A single request that has to hit 20 different services will have 20 spans associated with it. Those spans get collected into a trace. A trace can be evaluated to look at the latencies of each of those services.
If you are trying to improve the speed of a distributed systems infrastructure, distributed tracing can be very helpful for choosing where to focus your attention.
The published Google papers of ten years ago often turn out to be the companies of today. Some examples include MapReduce (which formed the basis of Cloudera), Spanner (which formed the basis of CockroachDB), and Dremel (which formed the basis of Dremio).
Today, a decade after he started thinking about distributed tracing, Ben Sigelman is the CEO of Lightstep, a company that provides distributed tracing and other monitoring technologies.
Lightstep’s distributed tracing model still bears a resemblance to the same techniques described in the paper–so I was eager to learn the differences between open source versions of distributed tracing (such as OpenZipkin) and enterprise providers such as Lightstep.
The key feature of Lightstep that we discussed: garbage collection.
If you are using a distributed tracing system, you could be collecting a lot of traces. You could collect a trace for every single user request. Not all of these traces are useful–but some of them are very useful. Maybe you only want to keep track of traces that take an exceptionally long latency. Maybe you want to keep every trace in the last 5 days, and destroy them over time. So, the question of how to manage the storage footprint of those traces was as interesting as the discussion of distributed tracing itself.
Beyond the distributed tracing features of his product, Ben has a vision for how his company can provide other observability tools over time. I spoke to Ben at Kubecon–and although this conversation does not talk about Kubernetes specifically, this topic is undoubtedly interesting to people who are building Kubernetes technologies.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
Docker was released in 2013, and popularized the use of containers. A container is an abstraction for isolating a well-defined portion of an operating system. Developers quickly latched onto containers as a way to cut down on the cost of virtual machines–as well as isolate code and simplify deployments. Developers began deploying so many containers that they needed a centralized way to manage the containers.
Then came the rise of the container orchestration framework.
A container orchestration framework is used to manage operating system resources. Twitter had been using the open source orchestration framework Apache Mesos to manage resources since 2010. So when developers started looking for a way to manage containers, many of them chose Mesos. Meanwhile, another container orchestration system came onto the scene: Docker Swarm. Swarm is a tool for managing containers that came from the same people who originally created Docker.
Shortly thereafter, Kubernetes came out of Google. Google had been using containers internally with their cluster manager Borg. Kubernetes is a container orchestration system that was built with the lessons of managing resources at Google.
The reason I recount this history (as I have in a few past episodes) is to underscore that there was a few years where there was a lot of debate around the best container orchestration system to use. Some people preferred Swarm, some preferred Mesos, and some preferred Kubernetes.
Because of the lack of standardization, the community of developers who were building infrastructure in this space were put in a tough position: should you pick a specific orchestration framework, and go all in, building only tools for that one framework? Should you try to build tools to be compatible with all three frameworks, and attempt to satisfy Kubernetes, Mesos, and Swarm users? Or should you sit out altogether and build nothing?
The fracturing of the community led to healthy debates, but it slowed down innovation, because different developers were moving in different directions. Today, the container community has centralized: Kubernetes has become the most popular container orchestration framework.
With the community centralizing on Kubernetes, developers are able to comfortably bet big on open source projects like Istio, Conduit, Rook, Fluentd, and Helm, each of which we will be covering in the next few weeks.
The centralization on Kubernetes also makes it easier to build enterprise companies, who are no longer trying to think about which container orchestration to support. There is a wide array of Kubernetes-as-a-service providers offering a highly available runtime–and a variety of companies offering observability tools to make it easier to debug distributed systems problems.
Despite all of these advances–Kubernetes is less usable than it should be. It still feels like operating a distributed system. Hopefully someday, operating a Kubernetes cluster will be as easy as operating your laptop computer. To get there, we need improvements in Kubernetes usability.
Today’s guest Joe Beda was one of the original creators of the Kubernetes project. He is a founder of Heptio, a company that provides Kubernetes tools and services for enterprises. I caught up with Joe at KubeCon 2017, and he told me about where Kubernetes is today, where it is going, and what he is building at Heptio. Full disclosure–Heptio is a sponsor of Software Engineering Daily.
For the next two weeks, we are covering exclusively the world of Kubernetes. Kubernetes is a project that is likely to have as much impact as Linux. Whether you are an expert in Kubernetes or you are just starting out, we have lots of episodes to fit your learning curve. To find all of our old episodes about Kubernetes, download the Software Engineering Daily app for iOS or for Android. In other podcast players, only the most 100 recent episodes are available, but in our apps you can find all 650 episodes–and there is also plenty of content that is totally unrelated to Kubernetes!
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
In the first 10 years of cloud computing, a set of technologies emerge that every software enterprise needs; continuous delivery, version control, logging, monitoring, routing, data warehousing. These tools were built into the Cloud Foundry project, a platform for application deployment and management.
As we enter the second decade of cloud computing, another new set of technologies are emerging as useful tools. Serverless functions allow for rapid scalability at a low cost. Kubernetes offers a control plane for containerized infrastructure. Reactive programming models and event sourcing make an application more responsive and simplify the interactions between teams who are sharing data sources.
The job of a cloud provider is to see new patterns in software development and offer tools to developers to help them implement those new patterns. Of course, building these tools is a huge investment. If you’re a cloud provider, your customers are trusting you with the health of their application. The tool that you build has to work properly and you have to help the customers figure out how to leverage the tool and resolve any breakages.
Onsi Fakhouri is the senior VP of R&D for cloud at Pivotal, a company that provides a software and support for Spring, Cloud Foundry and several other tools. I sat down with Onsi to discuss his strategy for determining which products Pivotal chooses to build. There are a multitude of engineering and business elements that Onsi has to consider when allocating resources to a project.
Cloud Foundry is used by giant corporations like banks, telcos and automotive manufacturers. Spring is used by most enterprises that run Java, including most of the startups that I have worked at in the past. Cloud Foundry has to be able to run on premise and in the cloud providers like AWS, Google and Microsoft. Pivotal also has its own cloud, Pivotal Web Services, and all of these stakeholders have different technologies that they would like to see built. Onsi’s job is to determine which ones have the highest net impact and make a decision on those and allocate resources towards them.
I interviewed Onsi at Spring One Platform, which is a conference that is organized by Pivotal who, full disclosure, is a sponsor of Software Engineering Daily. This week’s episodes are all conversations from that conference, and if there’s a conference that you think I should attend and do coverage at, let me know. Whether you like this format or not, I would love to get your feedback. We have some big developments coming for Software Engineering Daily in 2018 and we want to have a closer dialogue with the listeners. Please send me an email, [email protected] or join our Slack channel.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
In the 1980s and the 1990s, most applications used only a relational database for their data management. In the early 2000s, software projects started to use an ever increasing number of data sources. MongoDB popularized the document database, which allows storage of objects that do not have a consistent schema.
The Hadoop distributed file system enabled the redundant storage and efficient querying of high volumes of data that are spread out across multiple commodity disks. The Cassandra Database is a hybrid between key-value storage and column-oriented storage.
The benefit of these different data systems is that you can choose a system that gives you the read and write performance that you need. The downside is that each of these databases has different querying semantics. If you’re a developer trying to access data from your application, you often need to know how to access that data from the specific data source and whether that data needs to be queried with SQL, or with the document style query, or with a MapReduce job.
Spring Data is a project to standardize the programming model for data access within Spring. The vision for the project is to give Spring developers a consistent way to access their data from any database, or retaining the performance characteristics of those databases.
Spring is a Java framework for writing web applications, but this conversation is useful even for people who are not building these Spring applications. Whatever application you’re building, you are probably pulling from multiple data sources. The question of how to abstract away the complexity of those multiple data sources is also being tackled by projects such as GraphQL and Falcor.
John Blum is a staff engineer who works on the Spring Data Project at Pivotal. He joins the show to discuss how to design a data access layer. We discussed the API between a database and the Spring Data layer and also talked about reactive programming. Reactive programming allows the application layer to respond to changes in the underlying data layer.
I interviewed John at SpringOne Platform, which is a conference that is organized by Pivotal, who full disclosure is a sponsor of Software Engineering Daily. This week’s episodes are all conversations from that conference.
If there’s a conference that you think I should attend and do some coverage at, please let me know. Whether you like this format or not, I would love to get your feedback. We have some big developments coming for Software Engineering Daily in 2018, and we want to have a closer dialogue with the listeners. Please send me an e-mail [email protected], let me know what’s up. Or join our Slack channel.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
Cloud Foundry is an open-source platform as a service for deploying and managing web applications. Cloud Foundry is widely used by enterprises who are running applications that are built using Spring, a popular web framework for Java applications, but developers also use Cloud Foundry to manage apps built in Ruby, Node and any other programming language. Cloud Foundry includes routing, message brokering, service discovery, authentication and other application level tooling for building and managing a distributed system. Some of the standard tooling in Cloud Foundry was adopted from Netflix open-source projects, such as Hystrix, which is the circuit breaker system; and Eureka, which is the service discovery server and client.
When a developer deploys their application to Cloud Foundry, the details of what is going on are mostly abstracted away, which is by design. When you’re trying to ship code and iterate quickly for your organization, you don’t want to think about how your application image is being deployed to underlying infrastructure. You don’t want to think about whether you’re deploying a container or a VM, but if you use Cloud Foundry enough, you might have become curious about how Cloud Foundry schedules and runs application code.
BOSH is a component of Cloud Foundry that sits between the infrastructure layer and the application layer. Cloud Foundry can be deployed to any cloud provider because of BOSH’s well-defined interface. BOSH has the abstraction of a stem cell, which is a versioned operating system image wrapped in packaging for whatever infrastructure as a service is running underneath. With BOSH, whenever a VM gets deployed no your underlying infrastructure, that VM gets a BOSH agent. The agent communicates with the centralized component of BOSH called the director. This role of director is the leader of the distributed system.
Rupa Nandi is a director of engineering at Pivotal where she works on Cloud Foundry. In this episode we talked about scheduling an infrastructure, the relationship between Spring and Cloud Foundry and the impact of Kubernetes, which Cloud Foundry has integrated with so that users can run Kubernetes workloads on Cloud Foundry.
I interviewed Rupa at SpringOne Platform, a conference that is organized by Pivotal who, full disclosure, is a sponsor of Software Engineering Daily, and this week’s episode are all conversations from that conference. Whether you like this format or don’t like this format, I would love to get your feedback. We have some big developments coming for Software Engineering Daily in 2018 and we want to have a closer dialogue with the listeners. Please send me an email, [email protected] or join our Slack channel. We really want to know what you’re thinking and what your feedback is, what you would like to hear more about, what you’d like to hear less about, who you are.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
Originally published November 18, 2015
Seth Godin is a writer, speaker, and entrepreneur. He is the author of many books, including most recently, What To Do When It’s Your Turn.
Originally published November 17, 2015
Technical debt, referring to the compounding cost of changes to software architecture, can be especially challenging in machine learning systems.
D. Sculley is a software engineer at Google, focusing on machine learning, data mining, and information retrieval. He recently co-authored the paper Machine Learning: The High Interest Credit Card of Technical Debt.
Military force is powered by software.
The drones that are used to kill suspected terrorists can identify those terrorists using the same computer vision tools that are used to identify who is in an Instagram picture. Nuclear facilities in Iran were physically disabled by the military-sponsored Stuxnet virus. National intelligence data is collected and processed using the MapReduce algorithm.
The military keeps up with technology more effectively than lawmakers. It is common to read a quote from a senator or a judge that shows a basic misunderstanding of cybersecurity. Many politicians do not even use email.
There is a large and growing knowledge gap between military capability and the technological savvy of policymakers. On the whole, government is not prepared for modern warfare.
We are lucky that military conflict in 2017 is decentralized.
There are small skirmishes in the Middle East, but there is nothing compared to the centralized hostility of World War II. The bonds of international trade form some protection against a war between major powers like US and China.
But it is easy to imagine circumstances that could lead global war. There is an arms race between the United States and China—and this arms race is happening in an environment of increasing informational chaos.
Online journalism is becoming physically dangerous. Journalists are being tracked and threatened. Locations where journalists live have been published by enemies of those journalists. And this war on journalists has a chilling effect on the production of truthful information.
In the war between truth and false information, the people promoting false information have a giant advantage. And that advantage is growing because we cannot identify who is a bot and who is a human.
Social media creates a fog of war around information.
There is also a fog of war around real world physical conflict that is enacted remotely through computers. This type of confusion is the problem known as cybersecurity attribution.
When a power grid gets knocked offline by a hacker, how do we know who was responsible? When a drone flies through New York City shooting at people, how do we know who deployed that drone? When self-driving car technology becomes open source, how long will it be before we see a 9/11-type event performed remotely with cars instead of an airplane—and when it does happen, how will we respond?
Imagine the level of grief and anger on 9/11 that caused our government to launch a war on Iraq. Whether or not you supported that war, we were at least (sort of) aiming in the right direction of the geographic location where the attackers came from.
What would we do if could not figure out who launched such an attack?
When cyber forensics teams look into a problem, it might appear that an attack came from China at one layer, Russia from another layer, and some domestic attacker at another layer. It can be time consuming to uncover the truth behind a cyberattack.
Just like in social media information wars, the instigators of conflict have an advantage.
And the ability to instigate such a conflict is democratized. Social media, open source software, and cloud computing give a technologist superpowers. Cryptocurrencies can anonymize the financial transactions to pay for such tools, and basic encryption can anonymize the terroristic acts that occur over a remote internet connection.
Peter Warren Singer is a political scientist who formerly worked in the United States advisory committee on International Communications and Information Policy. He is also an author, whose books include Wired for War, Cybersecurity and Cyberwar: What Everyone Needs to Know, and Ghost Fleet: A Novel of the Next World War. Peter writes about the circumstances that could lead to global warfare, and how military actors might behave in a third world war.
In this episode, Peter shares a dark, but realistic vision that we should all hope to avoid.
If you like this episode, we have done many other shows on related topics–including drones, IoT security, and automotive cybersecurity. To find these old episodes, you can download the Software Engineering Daily app for iOS and for Android. In other podcast players, you can only access the most recent 100 episodes. With these apps, we are building a new way to consume content about software engineering. They are open-sourced at github.com/softwareengineeringdaily. If you are looking for an open source project to get involved with, we would love to get your help.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
The New York Times makes most of its money off of subscriptions. Facebook makes its money off of native advertising. Hacker News is funded by Y-Combinator. Each of these business models creates biases in the information that gets promoted on the respective platforms.
This is why I like to know the origin story and the business models behind the publications that I read. Published content is shaped by the profit motive of the publication.
And yet, last month, I repeatedly found myself reading high quality content on a Medium publication that I did not know the origin of: Hacker Noon. Hacker Noon is a popular Medium publication that syndicates curated content written about software.
Let me explain “syndication.” Imagine that I just spent three days on a Medium post about functional programming, and I have zero followers on social media. How can I get people to read awesome post? The answer is syndication. I can submit my Medium post to Hacker Noon. This gives me free distribution, and it gives Hacker Noon free content—a win-win relationship.
But why was it worth it for Hacker Noon to spend time curating content? That syndication process takes time. You have to read through lots of submissions, sometimes you have to send it back to the author to have it edited. And this is all to build a following on Medium. I have not heard of Medium being a profitable platform to build a business.
It’s worth pointing out the difference between Medium and WordPress.
On WordPress, this model of curated syndication has worked to massive success—for example, the Huffington Post and TechCrunch. These businesses make millions of dollars from advertising networks, because they are built on WordPress, and WordPress is an open model.
A publisher on WordPress can install plugins that serve ads from third party providers like Outbrain and Taboola. A WordPress site can also install any kind of data collection scripts, to gather data on visitors, and sell it to the highest bidder.
The lack of third party plugins is the blessing and the curse of Medium.
Because there is no third party ecosystem, reading content on Medium is a beautiful experience. The page loads quickly and predictably. There are no random scripts that are blocking the page as they hog your browser’s resources. When you go to close the page, there is never a popup that asks you to subscribe to a newsletter.
When I read content on Medium, I am not getting slapped across the face with ads for reverse mortgages and açaí berries. I am not being tagged for retargeting. It’s a beautiful experience. But Medium seems like an ecosystem that would not allow for the content syndication business like Hacker Noon.
I wanted to know who was running Hacker Noon, how the business works, and what it says about Medium as a publishing platform.
Hacker Noon turns out to be part of a network of Medium publications called AMI. AMI’s network includes sites like Art + Marketing, Future Travel, and Fit Yourself Club–all of which are distinct syndication platforms. David Smooke is the CEO of AMI, and he joins this episode to explain how his business works, how he has scaled the content syndication business, and why he is betting on Medium. It was a detailed look into the state of online publishing and where it might be headed.
If you don’t read Hacker Noon already, one article to start with that shows off the quality of content is Learn Blockchains By Building One. I interviewed the author of that article, Daniel Van Flymen, and it has been one of the most popular episodes of Software Engineering Daily.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
From the publisher's feed