Alternatives to Google Cloud Pub/Sub

The algorithms and data infrastructure at Stitch Fix is housed in #AWS. Data acquisition is split between events flowing through Kafka, and periodic snapshots of PostgreSQL DBs. We store data in an Amazon S3 based data warehouse. Apache Spark on Yarn is our tool of choice for data movement and #ETL. Because our storage layer (s3) is decoupled from our processing layer, we are able to scale our compute environment very elastically. We have several semi-permanent, autoscaling Yarn clusters running to serve our data processing needs. While the bulk of our compute infrastructure is dedicated to algorithmic processing, we also implemented Presto for adhoc queries and dashboards.

Beyond data movement and ETL, most #ML centric jobs (e.g. model training and execution) run in a similarly elastic environment as containers running Python and R code on Amazon EC2 Container Service clusters. The execution of batch jobs on top of ECS is managed by Flotilla, a service we built in house and open sourced (see https://github.com/stitchfix/flotilla-os).

At Stitch Fix, algorithmic integrations are pervasive across the business. We have dozens of data products actively integrated systems. That requires serving layer that is robust, agile, flexible, and allows for self-service. Models produced on Flotilla are packaged for deployment in production using Khan, another framework we've developed internally. Khan provides our data scientists the ability to quickly productionize those models they've developed with open source frameworks in Python 3 (e.g. PyTorch, sklearn), by automatically packaging them as Docker containers and deploying to Amazon ECS. This provides our data scientist a one-click method of getting from their algorithms to production. We then integrate those deployments into a service mesh, which allows us to A/B test various implementations in our product.

For more info:

• Our Algorithms Tour: https://algorithms-tour.stitchfix.com/
• Our blog: https://multithreaded.stitchfix.com/blog/
• Careers: https://multithreaded.stitchfix.com/careers/

#DataScience #DataStack #Data

As we've evolved or added additional infrastructure to our stack, we've biased towards managed services. Most new backing stores are Amazon RDS instances now. We do use self-managed PostgreSQL with TimescaleDB for time-series data—this is made HA with the use of Patroni and Consul.

We also use managed Amazon ElastiCache instances instead of spinning up Amazon EC2 instances to run Redis workloads, as well as shifting to Amazon Kinesis instead of Kafka.

As Sentry runs throughout the day, there are about 50 different offline tasks that we execute—anything from “process this event, pretty please” to “send all of these cool people some emails.” There are some that we execute once a day and some that execute thousands per second.

Managing this variety requires a reliably high-throughput message-passing technology. We use Celery's RabbitMQ implementation, and we stumbled upon a great feature called Federation that allows us to partition our task queue across any number of RabbitMQ servers and gives us the confidence that, if any single server gets backlogged, others will pitch in and distribute some of the backlogged tasks to their consumers.

#MessageQueue

Around the time of their Series A, Pinterest’s stack included Python and Django, with Tornado and Node.js as web servers. Memcached / Membase and Redis handled caching, with RabbitMQ handling queueing. Nginx, HAproxy and Varnish managed static-delivery and load-balancing, with persistent data storage handled by MySQL.

Hi Otensia! I'd definitely recommend using the skills you've already got and building with JavaScript is a smart way to go these days. Most platform services have JavaScript/Node SDKs or NPM packages, many serverless platforms support Node in case you need to write any backend logic, and JavaScript is incredibly popular - meaning it will be easy to hire for, should you ever need to.

My advice would be "don't reinvent the wheel". If you already have a skill set that will work well to solve the problem at hand, and you don't need it for any other projects, don't spend the time jumping into a new language. If you're looking for an excuse to learn something new, it would be better to invest that time in learning a new platform/tool that compliments your knowledge of JavaScript. For this project, I might recommend using Netlify, Vercel, or Google Firebase to quickly and easily deploy your web app. If you need to add user authentication, there are great examples out there for Firebase Authentication, Auth0, or even Magic (a newcomer on the Auth scene, but very user friendly). All of these services work very well with a JavaScript-based application.

This is my stack in Application & Data

JavaScript PHP HTML5 jQuery Redis Amazon EC2 Ubuntu Sass Vue.js Firebase Laravel Lumen Amazon RDS GraphQL MariaDB

My Utilities Tools

Google Analytics Postman Elasticsearch

My Devops Tools

Git GitHub GitLab npm Visual Studio Code Kibana Sentry BrowserStack

My Business Tools

Slack

I use Socket.IO because the application has 2 frontend clients, which need to communicate in real-time. The backend-server handles the communication between these two clients via websockets. Socket.io is very easy to set up in Node.js and ExpressJS.

In the research project, the 1st client shows panoramic videos in a so called cave system (it is the VR setup of our research lab, which consists of three big screens, which are specially arranged, so the user experience the videos more immersive), the 2nd client controls the videos/locations of the 1st client.

We are starting to work on a web-based platform aiming to connect artists (clients) and professional freelancers (service providers). In-app, timeline-based, real-time communication between users (& storing it), file transfers, and push notifications are essential core features. We are considering using Node.js, ExpressJS, React, MongoDB stack with Socket.IO & Apollo, or maybe using Real-Time Database and functionalities of Firebase.

Which messaging service (Pusher vs. PubNub vs. Google Cloud Pub/Sub) to use for IoT?

Recently we finished long research on chat tool for our students and mentors. In the end we picked Mattermost Team Edition as the cheapest and most feature complete option. We did consider building everything from scratch and use something like Pusher or Twilio on a backend, but then we would have to implement all the desktop and mobile clients and all the features oursevles. Mattermost gave us flexible API, lots of built in or easy to install integrations and future-proof feature set. We are still integrating it with our main platform but so far the team, existing mentors and students are very happy.

We need to interact from several different Web applications (remote) to a client-side application (.exe in .NET Framework, Windows.Console under our controlled environment). From the web applications, we need to send and receive data and invoke methods to client-side .exe on javascript events like users onclick. SignalR is one of the .Net alternatives to do that, but it adds overhead for what we need. Is it better to add SignalR at both client-side application and remote web application, or use gRPC as it sounds lightest and is multilingual?

SignalR or gRPC are always sending and receiving data on the client-side (from browser to .exe and back to browser). And web application is used for graphical visualization of data to the user. There is no need for local .exe to send or interact with remote web API. Which architecture or framework do you suggest to use in this case?

I want to use NATS for my IoT Platform and replace it instead of the MQTT broker. is there any preferred added value to do that?