Stockage Central et Se curise De Donne es Vide o Isilon SCALE-OUT NAS bertrand.ounanian@emc.com 1
Introduction Stockage et Production Vidéo Vidéo et Monde Sportif Vidéo et Traitements Analytiques Applications à la Santé 2
Création de contenu vidéo Stations d édition Live Serveurs Ingest/Playout Stockage dédié SAN/DAS MAM Serveur d archive Archive NearLine NAS Archive Bande 3 3
Spécificités du Stockage Vidéo Perfomance Séquentialité des traitements Traitements non-linéaires Evolutivité Fiabilité résilience Ouverture Simplicité! 4
Simplification des workflows Stations d édition Live Ingest/Playout Servers Production Near-line & Archive MAM Archive Server Copyright 2013 2011 EMC EMC Corporation. All rights All reserved. rights reserved. 5
Content Delivery Media Flow Overview Broadcast Transmitter HiRes Media Source Prepare Media Content Management Conditional Access Control Originate Media Deliver Media Present/ Consume Media Mezzanine Quality Licensed Movies Live Programming Media Storage Transcode Segment Clone Media Storage MAM DAM Encrypt DRM File Streaming Origin Storage Cache Servers WebDAV CDN Brokers Cache Servers Web Servers Edge Storage Application Brokers STB PC/Mac Tablet Smartphone BYOD Internet Transmitter 6
Content Delivery CDN Overview 7
Example model IPTV Streaming IPTV Streaming 0.8 million subs CDN Sizing worksheet Originate Media Deliver Media Present/ Consume Media 10% concurrency File Streaming Origin Storage Cache Servers WebDAV CDN Brokers Cache Servers Web Servers Edge Storage Application Brokers STB PC/Mac Tablet Smartphone BYOD Average Bitrate 2.0 Mbps 1 Origin 5 Locations 23 POPs 116 streams 98% Cache Hit 1,163 streams 85% Cache Hit 0.2 Gb/s 2.3 Gb/s 3,370 streams 253 6.7 Gb/s 0.84 GB/s 253 Origin Server Distribution Servers Edge Cache Servers VOD + CatchUp TV Streams VOD + CatchUp TV Streams LIVE + VOD + CatchUp TV Streams 5 X400 3 X400 3 S200 393 TiB useable 154 TiB useable 21 TiB useable 8
Sport and Video uses Live & Entertainment Technique Analysis Team Performance Coaching Security 9
Video Surveillance Analytics Examples Retail Loitering & Traffic Analysis Perimeter Trip Wire People Counting Analytic Description Example Use Case Motion Detection When pixels change in a scene, the system will denote this. Only record on motion scenarios are prevalent such that recording for a camera is only when there is motion plus 60 seconds before and after motion starts/stops. Object Left Behind Detection of when a person drops a bag or leaves a bag in a scene. In train stations and transportation hubs, this is common to identify potential bomb threats. Trip Wire Detection of a person cross a line (directionally or holistically) in a field of view. At airports, a trip wire is used to detect someone going the wrong direction at an exit door from arrivals to baggage claim. Detect person crossing fence line, such as a border. People Counting Counting the number of people crossing a set line. In a entry way of a retail store, getting an idea of how many people are entering store over certain time period. Piggy Backing Identifying if more than 1 person enters a building at a time when only 1 badge swipe is used. Used in building management scenarios in conjunction with the access control system. Genetec and Next Level are good examples. License Plate Identifying and exporting the license plate numbers of a car. Use typically at main gates of secure facilities or in transportation scenarios such as highways to help police identify where a particular suspect might be. Facial Recognition Analytics tied to identify a person or just pass the facial information to another system. High security areas where identification of certain suspects is prevalent. Casinos using this to identify known watch list entities. 10
Quality of Care Deliver Better Health Care Big Data Analytics Accelerates the Path to Health Care HIGH Legacy System Database BI Reporting Big Data Analytics Delivering 10 Years Of Data In Seconds Associative Rule Mining and User Clustering Improves Pathways External Data Sources Enable Personalized Medicine LOW Treatment Pathways on Summary Data Treatment Pathways on All the Data TRADITIONAL DATA LEVERAGED BIG DATA LEVERAGED 11
Imaging Environment High Level Overview Modality/Device Viewing Workstation Application Database Image Cache 1 2 CT Scans Ultrasound Mammography MRI etc. 3 Application Server Image Archive DR Image Archive 1 2 3 High Level Workflow Explanation Image captured from modality and sent to Application Server Viewing workstations access Application Server for output from modality (MRI, CT Scan etc) PACS Application updates DB, and then writes Image to Image Cache and Image Archive CDICOM) 12
Typical PACS Environment Image Viewer Image Viewer Image Viewer Image Viewer Facility A PACS Facility B PACS Cardiology Imaging Other Imaging (Ophth / Endo / Derm) Other Modalities Other Modalities Image Storage Image Storage Image Storage Image Storage 13
Vendor Neutral Archive (VNA) Environment Unified Image Viewer Facility A PACS Facility B PACS Cardiology Imaging Other Imaging (Ophth / Endo / Derm) Other Modalities Other Modalities Central Vendor Neutral Archive 14
Mutualisation des Données de Recherche Remote Data Acquisition (Sequencing Service Provider) cloud compute cloud storage Instruments onsite Aspera Tier 1 Storage Isilon S and X series HPC HPC PHI A ARCHIVE Annotation DB Experimental Data Tier 2 Storage Isilon NL series VIDEO SURVEILANCE ANALYTICS HOME DIR Tier 2 Storage Isilon NL series Local researchers and Analysts SYNCPLICIY PANORAMA Collaborators 15
Architecture en Cluster : Simplification Servers Servers 1GbE 10GbE 1GbE 10GbE Infiniband Servers Client/Application Layer Ethernet Layer Clustered Storage Layer Intra-cluster Communication Layer 16
Centralisation du Stockage Module 4: Horizontal and Vertical Markets 18
Poles universitaires / CHU elearning Production Media Animation Archive Surveillance 19
Bertrand Ounanian bertrand.ounanian@emc.com Benjamin Coutière benjamin.coutiere@emc.com 07 86 87 38 53 20