Identify, analyze and report networking issues in streaming services
Date & Time
Tuesday, November 10, 2020, 9:00 PM - 9:30 PM
Pranav Singh Horia-Mihai Popa Jason Thibeault


End-user streaming services have grown exponentially in the past years. The mechanisms to serve the enormous amounts of content available have seen a lot of improvements in terms of reliability, flexibility, and overall quality. However, significant problems that affect the content quality exist in the last -segment of the delivery infrastructure: home networks. The home networks -are oftern unmanaged and poor configuration and misuse - directly affect the quality of the video streaming experience. Detecting home network issues and appropriately reporting them is key to being able to provide - appropriate remedial solutions when users report problems. We propose a Microservices driven system that will cover these aspects: 1. Client-side services that collects data about the local network and reports issues - 2. Cloud-enabled series of microservices that: a. Collect data from the clients b. Store the data collected from the clients c. Analyze the data and create summarized information d. Send alerts into various systems (to allow preemptive action or on-demand action) The client-side library should be available on a variety of platforms: Windows, OS X, Linux, iOS, Android, WebOS, Tizen, etc. The cloud-side should provide: 1. Support for multi-tenancy 2. Ability to store data securely (on a per-tenant basis, or in a “public pool”) 3. Various levels of analysis on the data: a. Basic (just storage and the ability to access it) b. Structured (storage + organize the collected data to be more easily digested) c. Pre-analyzed (structured + basic analysis on the structured data) i. Provide information about the frequency with which some events happen ii. Basic correlation between various events (ex: event X happens if event Y is present) d. Fully analyzed data (pre-analyzed + predictions) i. Create predictions related to a specific issue (or set of issues) ii. Create predictions on a per-user basis (for the issues that are likely to happen to that specific user)

Technical Depth of Presentation
The technical depth will be at an intermediate level to ensure a wider understanding of the audience
What Attendees will Benefit Most from this Presentation
C-Suite Executives, Vice Presidents, Senior Manager, Technologists involved in Video Delivery The ideal audience would be: • Operation department executives (because they are the ones that need to solve the issues that are happening; our solution would provide them the appropriate information when dealing with the problems raised by the users) • Technology leads (CTOs, VPs of engineering, etc. – they need to understand how this can be integrated in existing and/or new products) • Product managers (Exec level – the issues that we plan on tackling are issues that ultimately affect the churn rate, and we can improve that)
Take-Aways from this Presentation
• Add last-segment of the delivery infrastructure insights & improve issue mitigation options • Improve service reliability & user experience • Improve the ability of the support teams to respond to user issues