Articles via Databases
Articles via Journals
Online Catalog
E-books
Research & Information Literacy
Interlibrary loan
Theses & Dissertations
Collections
Policies
Services
About / Contact Us
Administration
Littman Architecture Library
This site will be removed in January 2019, please change your bookmarks.
This page will redirect to https://digitalcommons.njit.edu/dissertations/1478/ in 5 seconds

The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title: Live media production: multicast optimization and visibility for clos fabric in media data centers
Author: Latif, Ammar
View Online: njit-etd2020-033
(xiii, 96 pages ~ 10.7 MB pdf)
Department: Department of Electrical and Computer Engineering
Degree: Doctor of Philosophy
Program: Electrical Engineering
Document Type: Dissertation
Advisory Committee: Khreishah, Abdallah (Committee chair)
Ansari, Nirwan (Committee member)
Borcea, Cristian (Committee member)
Liu, Qing Gary (Committee member)
Rayes, Amma (Committee member)
Date: 2020-08
Keywords: Controller
Live media production
Machine learning
Multicast
SDN
Video quality
Availability: Unrestricted
Abstract:

Media production data centers are undergoing a major architectural shift to introduce digitization concepts to media creation and media processing workflows. Content companies such as NBC Universal, CBS/Viacom and Disney are modernizing their workflows to take advantage of the flexibility of IP and virtualization.

In these new environments, multicast is utilized to provide point-to-multi-point communications. In order to build point-to-multi-point trees, Multicast has an established set of control protocols such as IGMP and PIM. The existing multicast protocols do not optimize multicast tree formation for maximizing network throughput which lead to decreased fabric utilization and decreased total number of admitted flows. In addition, existing multicast protocols are not bandwidth-aware and could cause links to over-subscribe leading to packet loss and lower video quality.

TV production traffic patterns are unique due to ultra high bandwidth requirements and high sensitivity to packet loss that leads to video impairments. In such environments, operators need monitoring tools that are able to proactively monitor video flows and provide actionable alerts. Existing network monitoring tools are inadequate because they are reactive by design and perform generic monitoring of flows with no insights into video domain.

The first part of this dissertation includes a design and implementation of a novel Intelligent Rendezvous Point algorithm iRP for bandwidth-aware multicast routing in media DC fabrics. iRP utilizes a controller-based architecture to optimize multicast tree formation and to increase bandwidth availability in the fabric. The system offers up to 50\% increase in fabric capacity to handle multicast flows passing through the fabric.

In the second part of this dissertation, DiRP algorithm is presented. DiRP is based on a distributed decision-making approach to achieve multicast tree capacity optimization while maintaining low multicast tree setup time. DiRP algorithm is tested using commercially available data center switches. DiRP algorithm offers substantially lower path setup time compared to centralized systems while maintaining bandwidth awareness when setting up the fabric.

The third part of this dissertation studies the utilization of machine learning algorithms to improve on multicast efficiency in the fabric. The work includes implementation and testing of LiRP algorithm to increase iRP's fabric efficiency by implementing k-fold cross validation method to predict future multicast group memberships for time-series analysis. Testing results confirm that LiRP system increases the efficiency of iRP by up to 40\% through prediction of multicast group memberships with online arrival.

In the fourth part of this dissertation, The problem of live video monitoring is studied. Existing network monitoring tools are either reactive by design or perform generic monitoring of flows with no insights into video domain. MediaFlow is a robust system for active network monitoring and reporting of video quality for thousands of flows simultaneously using a fraction of the cost of traditional monitoring solutions. MediaFlow is able to detect and report on integrity of video flows at a granularity of 100 mSec at line rate for thousands of flows. The system increases video monitoring scale by a thousand-fold compared to edge monitoring solutions.


If you have any questions please contact the ETD Team, libetd@njit.edu.

 
ETD Information
Digital Commons @ NJIT
Theses and DIssertations
ETD Policies & Procedures
ETD FAQ's
ETD home

Request a Scan
NDLTD

NJIT's ETD project was given an ACRL/NJ Technology Innovation Honorable Mention Award in spring 2003