Evan Langlais

Software Engineer II -- Azure Networking (SDN) on the Virtual Networking (VNET) team

Engineer working on new Accelerated Connection feature and platform Accelerated Connections MSDN

Software Engineer -- Productivity Suite Office Integration Team

Worked on collection of products to accelerate productivity of Investment Bankers using Excel, PowerPoint, and Word Office applications

Abstract

For security, isolation, metering and other purposes, public clouds today implement complex network functions at every server. Today's implementations, in software or on FPGAs and ASICs that are attached to each host, are becoming increasingly complex, costly and bottlenecks to scalability.

We present a different design that disaggregates network function processing off the host and into shared resource pools by making novel use of appliances which tightly integrate general-purpose ARM cores with high-speed stateful match processing ASICs. When work is skewed across VMs, such disaggregation can offer better reliability and performance over the state-of-art at a lower per-server cost.

We describe our solutions to the consequent challenges and present results from a production deployment at a large public cloud.

Presentation

• NSDI Page PDF Slides

In a way a spiritual successor to Treehouse Technologies, Xynon Ventures seeks to be a modern outlet for my ideas in the tech and non tech spaces alike.

The HuskyTHON Application is a modern web application that allows participants of the HuskyTHON Dance Marathon event at the University of Connecticut to keep up to date with various fundraising events, personal and team fundraising progress, and information regarding the dance marathon.

The application is a Vue3 web app utilizing the Ionic component library to create a cohesive 'native-esk' experience.

It also utilizes a Node.js App (hosted in Azure) which is responsible for brokering data from various sources into a single highly available and scalable API for the application.

• Frameworks: Vue 3
• Languages: Typescript, Javascript
• Platforms: iOS
• Technologies: Ionic 5, Capacitor

• Download in App Store
• View Privacy Policy (old)
• View Privacy Policy

Led a team of 6 academic peers to fulfill client requirement of allowing technicians to view data of, and send commands to, an existing infrastructure of over 7000 generators throughout the northeast, via a web app.

I became the project's lead architect, and after reading IOT and big-data white papers provided by our academic advisor I planned a five stage pipeline which would theoretically allow a full rollout of retrofitting the 7000+ generators, fully supporting the amount of data processing that would be associated with it. To implement the project, the team worked together to plan all necessary tasks, those tasks' dependencies, and estimated amount of work. Together, we were able to create a fully working prototype several weeks before UConn's Demo Day 2019.

The five stage data pipeline included

1. Attaching a Raspberry Pi to the generator's computer, running a Go app to communicate via the MODBUS protocol to read sensor data and change generator settings. Data was packaged into efficient structures via Arvo and sent to gateway server
2. The Go gateway server/router served several purposes, the main two being 1) keeping tabs on generator uptime by issuing heartbeats and 2) route incoming data packages to the processing pipeline and outgoing commands to the proper generator connection
3. The processing pipeline started with a Kafka cluster to ensure scalability, and data was consumed via a Scala app that would spun up in a Spark cluster. The Scala app would unpack the data and transform the input into time-series data, POSTing that to an OpenTSDB database
4. An express.js REST API was utilized to return generator inventory from MariaDB and their respective statuses from the gateway server, sensor time-series data, and relay commands back to the generator
5. An Angular web-app provided a dashboard for a technician to log in, view the health of all generators they have permission to view, and visualize sensor data over time utilizing chart.js

• Code Repository
• Technical Overview

HUE is a simple color matching game that was in development for the better part of the last two years, starting as a spiritual successor from a different original game called Occasionally Blue. The player uses the full screen to intuitively select a color that is compared against a set of shapes randomly drawn on the screen utilizing the Delta-E 2000 algorithm.

• Frameworks: Phaser.io
• Languages: Javascript
• Platforms: iOS and Android
• Technologies: Cordova, Firebase

• Download in App Store

Started as a personal project all the way back in 2014, the new version of the TF2 Finance app is a modern web application that utilizes Backpack.tf's currency API to display information and graphs regarding TF2 items such as Keys, Refined Metal, and Earbuds. Other features include converting currencies into other forms (including USD) and graphing price history for the application.

It also utilizes a PHP backend with caching to create a basic API for brokering information from various sources.

• Frameworks: Ionic 4, Angular 6
• Languages: Typescript, HTML, CSS, PHP
• Platforms: iOS and Android
• Technologies: Cordova, Firebase, AdMob

Treehouse Technologies was originally founded by myself and two friends with a wildly ambitious vision of building video games to compete with AAA game studios. After the other two founding members left, I registered Treehouse Technologies as a sole proprietorship in my freshman year of high-school. From that point, the Treehouse Technologies brand became one of self and business experimentation--constantly exploring new technology and business ventures.

I credit most of the problem solving and programming skills that I use today (both personally and professionally) to be a result of the creative experimentation that Treehouse Technologies enabled.