SlideShare une entreprise Scribd logo
1  sur  24
Не так страшен
терабит
Vyacheslav Olhovchenkov
Integros
Profile
• Full stack video solution
• Adaptive multibitrate streaming
• HLS, DASH & mp4 support, up to 1080p, native playback!
• Advertising standards supported: IMA, VPAID, VAST 2,3
• User platforms: desktop, mobile, STB, TV
• Support embedding into application
• Regions: CIS, EU, US, South America
• Content: UGC, advertising, professional
2
Load profile
• Peak bandwidth 1Tbit/s (200Gbit/s @ 2015)
• Peak bandwidth per edge 74Gbit/s (40Gbit/s @2015)
• Daily traffic 6+PTB
• Views per day 40M
• Aggregated storage cap: 1.3PB (0.7PB used)
3
What’s new?
• Moved to html5 player
• Dropped Flash & HDS support
• SSL support
• P2P in production
4
Core components
5
Edge (features)
• Cost-effectively, high-performance
• Store and delivery content to user
• Check URL signature
• Collect per-client stats
• Collect per-stream stats
• 60+K simulations users
• 74Gb/s output
6
Edge (architecture)
• FreeBSD 11
ZFS ZMIRROR
32TB HDD
ZFS L2ARC
3.5TB SSD
ZFS ARC
190GB
RAM
Nginx,
Lua
NIC
2x40Gbit
Internet
7
Edges history
• V1 (20Gbit), 2012
• V2 (42Gbit), 2013
• V3 (52Gbit), 2015
• V4 (74Gbit), 2016
8
Edge v4.0 (configuration)
• Dual E5-2650 v4 @ 2.20GHz
• 128GB RAM
• 8 SSD 500GB
• 18 HDD 4+TB
• DualPort Chelsio T580-LP-CR
Fail: 54.55Gbit/s limited (CPU idle: 28%)
9
Edge PCIe limitation
PCIe Speed Encoding Total (8x)
v1.0 2.5 Gb/s 8B/10B 16 Gbit/s
v2.0 5.0 Gb/s 8B/10B 32 Gbit/s
v3.0 8.0 Gb/s 128B/130B 63 Gbit/s
Gen PHY DLL ECRC
(opt)
TLP
header
Total Data 128 Data 256
1/2 1+1 2+4 4 16 24/28 .84/.82 .91/.90
3.0 4 2+4 4 16 26/30 .83/.81 .91/.90
TLP overhead (64bit):
Total: 8*8/130*128*256/(256+30) = 56.4
10
Edge v4.1 (configuration)
• Dual E5-2650 v4 @ 2.20GHz
• 256GB RAM
• 10 SSD 500GB
• 18 HDD 4+TB
• DualPort Chelsio T62100-LP-CR / Solarflare SFC9200
• Solarflare: firmware QSFP+ issues, BIOS bugs
• Chelsio: firmware uneven distributed, link down
11
Edge load profile
74 Gbit/s
12
Not all SSD are equally useful
• Crucial MX100 (ok)
• Crucial BX100 (ok)
• Crucial MX200 (medium speed ok)
• Crucial BX200
• Crucial MX300
• Fail at high speed (about ~35Gbit).
13
SSD comparison
MX100 MX200 BX200 MX300 850Pro 850Evo
read 475 482 463 456 506 495
write 443 335 232 310 457 391
mixed 423 326 248 318 478 398
read4k 24.8 24.4 25.2 21.6 32.4 38
write4k 83.3 68 43.9 60.2 57.6 77.9
mixed4k 24.6 19.9 17 17.8 34.7 35
-15% -33% -23% +17% +12%
14
SSL: performance drop
• Expectations (from Google): no performance drop.
• Reality: performance drop, load rise about 1% CPU load per Gigabit
(0.72GHz/Gbit => 1.22GHz/Gbit)
• Measurements (per Gigabit):
• SSL off: 1.11% sys + 0.25% user
• SSL on: 0.75% sys + 1.55% user
• 0.72 busy @ 33Gbit (Dual E5-2650v4 @ 2.20GHz), expected: 45Gbit
15
P2P (WebRTC)
• Goals:
• cost reduce
• QoS for distant sites
• Streamroot
• Hot content
• Up to 70% offload during spikes
• 90% for top 50 streams
16
P2P Architecture
1. Initialization of the video
2. Connection to the tracker
3. Direct exchanges with the network of viewers
1.
2.
3.
Pool of available
viewers
17
P2P Architecture
• Viewer authenticates and initializes the connection with Integros load
balancer
• At the same time, viewer connects to Streamroot tracker and
requests available peers
• Tracker sends back intelligently-selected list of peers that is updated
throughout the session according to bitrate and other criteria
• A P2P WebRTC DataChannel is created directly between viewers to
exchange segments
18
P2P wins
• Server offload
• Up to 70% all week
• 277Gbps offload on peak, 241 on 95/5 !!
• Live streams
• Start at 23 concurent viewers (50+% offloads)
19
Metric P2P ON P2P OFF Difference
Buffering rate 0.84% 1.05% -20%
Average Bitrate 1.08 Mbps 1.06 Mbps +1.9%
Session duration avg 1088s 1065s +2.2%
P2P wins (QoS)
• Doesn’t add ANY latency, for VoD or Live!
• The user experience is EXACTLY the same
20
P2P Issues and Current Status
• Device and browser support (80-90%% devices supported)
• Desktop
• Chrome, FF, Opera Yandex OK
• Safari 11+ OK since September !!
• Edge: OK
• Mobile
• Webview
• Chrome Android Ok
• Safari iOS11: soon
• Apps
• Android SDK OK (Exoplayer, Nexplayer, etc)
• iOS SDK OK (AVPlayer & Webview)
• STBs & SmartTVs:
• Android TV OK
• Apple TV soon
• Smart TV: not yet 
21
P2P Future
• Compatibility growth
• STBs
• Smart TVs!
• Continue R&D
• Better QoS algos
• Better P2P algos
22
Total traffic
23
Links
• http://ssd.userbenchmark.com/
• https://www.xilinx.com/support/documentation/white_papers/wp350.pdf
Questions?
slawa@integros.com
24

Contenu connexe

Tendances

Tendances (20)

Quality of Service Ingress Rate Limiting and OVS Hardware Offloads
Quality of Service Ingress Rate Limiting and OVS Hardware OffloadsQuality of Service Ingress Rate Limiting and OVS Hardware Offloads
Quality of Service Ingress Rate Limiting and OVS Hardware Offloads
 
Introduction to QUIC
Introduction to QUICIntroduction to QUIC
Introduction to QUIC
 
ReplacingSquidWithATS
ReplacingSquidWithATSReplacingSquidWithATS
ReplacingSquidWithATS
 
TRex Realistic Traffic Generator - Stateless support
TRex  Realistic Traffic Generator  - Stateless support TRex  Realistic Traffic Generator  - Stateless support
TRex Realistic Traffic Generator - Stateless support
 
Include os @ flossuk 2018
Include os @ flossuk 2018Include os @ flossuk 2018
Include os @ flossuk 2018
 
Load Balancing MySQL with HAProxy - Slides
Load Balancing MySQL with HAProxy - SlidesLoad Balancing MySQL with HAProxy - Slides
Load Balancing MySQL with HAProxy - Slides
 
Varnish Cache 4.0 / Redpill Linpro breakfast in Oslo
Varnish Cache 4.0 / Redpill Linpro breakfast in OsloVarnish Cache 4.0 / Redpill Linpro breakfast in Oslo
Varnish Cache 4.0 / Redpill Linpro breakfast in Oslo
 
Open vSwitch Implementation Options
Open vSwitch Implementation Options Open vSwitch Implementation Options
Open vSwitch Implementation Options
 
Geneve
GeneveGeneve
Geneve
 
Peter Zaitsev "18 ways to fix MySQL bottlenecks"
Peter Zaitsev "18 ways to fix MySQL bottlenecks"Peter Zaitsev "18 ways to fix MySQL bottlenecks"
Peter Zaitsev "18 ways to fix MySQL bottlenecks"
 
Openv switchの使い方とか
Openv switchの使い方とかOpenv switchの使い方とか
Openv switchの使い方とか
 
Memory, Big Data, NoSQL and Virtualization
Memory, Big Data, NoSQL and VirtualizationMemory, Big Data, NoSQL and Virtualization
Memory, Big Data, NoSQL and Virtualization
 
SPDY & HTTP2.0 & QUIC - #bpstudy 2013-08-28
SPDY & HTTP2.0 & QUIC - #bpstudy 2013-08-28SPDY & HTTP2.0 & QUIC - #bpstudy 2013-08-28
SPDY & HTTP2.0 & QUIC - #bpstudy 2013-08-28
 
Cumulus Linux 2.5.4
Cumulus Linux 2.5.4Cumulus Linux 2.5.4
Cumulus Linux 2.5.4
 
Open Networking for Your OpenStack
Open Networking for Your OpenStackOpen Networking for Your OpenStack
Open Networking for Your OpenStack
 
Managing Open vSwitch Across a Large Heterogenous Fleet
Managing Open vSwitch Across a Large Heterogenous FleetManaging Open vSwitch Across a Large Heterogenous Fleet
Managing Open vSwitch Across a Large Heterogenous Fleet
 
I want the next generation web here SPDY QUIC
I want the next generation web here SPDY QUICI want the next generation web here SPDY QUIC
I want the next generation web here SPDY QUIC
 
Open vSwitch Introduction
Open vSwitch IntroductionOpen vSwitch Introduction
Open vSwitch Introduction
 
Cassandra Performance Benchmark
Cassandra Performance BenchmarkCassandra Performance Benchmark
Cassandra Performance Benchmark
 
Protect your edge BGP security made simple
Protect your edge BGP security made simpleProtect your edge BGP security made simple
Protect your edge BGP security made simple
 

Similaire à Не так страшен терабит / Вячеслав Ольховченков (Integros)

(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
Amazon Web Services
 
Media Source Extensions
Media Source ExtensionsMedia Source Extensions
Media Source Extensions
FITC
 
Atf 3 q15-2 - product preview
Atf 3 q15-2 - product previewAtf 3 q15-2 - product preview
Atf 3 q15-2 - product preview
Mason Mei
 

Similaire à Не так страшен терабит / Вячеслав Ольховченков (Integros) (20)

Как построить видеоплатформу на 200 Гбитс / Ольховченков Вячеслав (Integros)
Как построить видеоплатформу на 200 Гбитс / Ольховченков Вячеслав (Integros)Как построить видеоплатформу на 200 Гбитс / Ольховченков Вячеслав (Integros)
Как построить видеоплатформу на 200 Гбитс / Ольховченков Вячеслав (Integros)
 
(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
(MED305) Achieving Consistently High Throughput for Very Large Data Transfers...
 
AWS re:Invent - Med305 Achieving consistently high throughput for very large ...
AWS re:Invent - Med305 Achieving consistently high throughput for very large ...AWS re:Invent - Med305 Achieving consistently high throughput for very large ...
AWS re:Invent - Med305 Achieving consistently high throughput for very large ...
 
High-performance 32G Fibre Channel Module on MDS 9700 Directors:
High-performance 32G Fibre Channel Module on MDS 9700 Directors:High-performance 32G Fibre Channel Module on MDS 9700 Directors:
High-performance 32G Fibre Channel Module on MDS 9700 Directors:
 
.NET Conf 2022 - Networking in .NET 7
.NET Conf 2022 - Networking in .NET 7.NET Conf 2022 - Networking in .NET 7
.NET Conf 2022 - Networking in .NET 7
 
Nokta techpresentation
Nokta techpresentationNokta techpresentation
Nokta techpresentation
 
Streaming video to html
Streaming video to htmlStreaming video to html
Streaming video to html
 
Развитие MX маршрутизаторов
Развитие MX маршрутизаторовРазвитие MX маршрутизаторов
Развитие MX маршрутизаторов
 
Azure VM 101 - HomeGen by CloudGen Verona - Marco Obinu
Azure VM 101 - HomeGen by CloudGen Verona - Marco ObinuAzure VM 101 - HomeGen by CloudGen Verona - Marco Obinu
Azure VM 101 - HomeGen by CloudGen Verona - Marco Obinu
 
PLNOG14: Konwergentność, Wydajność, Szybkość w Data Center - Kazimierz Jantas
PLNOG14: Konwergentność, Wydajność, Szybkość w Data Center - Kazimierz JantasPLNOG14: Konwergentność, Wydajność, Szybkość w Data Center - Kazimierz Jantas
PLNOG14: Konwergentność, Wydajność, Szybkość w Data Center - Kazimierz Jantas
 
Building a Router
Building a RouterBuilding a Router
Building a Router
 
BGP Flowspec (RFC5575) Case study and Discussion
BGP Flowspec (RFC5575) Case study and DiscussionBGP Flowspec (RFC5575) Case study and Discussion
BGP Flowspec (RFC5575) Case study and Discussion
 
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
 
IPv4 IPv6 Multi Protocol Media Player
IPv4 IPv6 Multi  Protocol Media PlayerIPv4 IPv6 Multi  Protocol Media Player
IPv4 IPv6 Multi Protocol Media Player
 
TV-IP262PI(v1.0R).pdf
TV-IP262PI(v1.0R).pdfTV-IP262PI(v1.0R).pdf
TV-IP262PI(v1.0R).pdf
 
Media Source Extensions
Media Source ExtensionsMedia Source Extensions
Media Source Extensions
 
IPv4 IPv6 Media Player
IPv4 IPv6 Media PlayerIPv4 IPv6 Media Player
IPv4 IPv6 Media Player
 
Atf 3 q15-2 - product preview
Atf 3 q15-2 - product previewAtf 3 q15-2 - product preview
Atf 3 q15-2 - product preview
 
Tacc Infinite Memory Engine
Tacc Infinite Memory EngineTacc Infinite Memory Engine
Tacc Infinite Memory Engine
 
Big Data Hadoop Briefing Hosted by Cisco, WWT and MapR: Cisco UCS For Big Dat...
Big Data Hadoop Briefing Hosted by Cisco, WWT and MapR: Cisco UCS For Big Dat...Big Data Hadoop Briefing Hosted by Cisco, WWT and MapR: Cisco UCS For Big Dat...
Big Data Hadoop Briefing Hosted by Cisco, WWT and MapR: Cisco UCS For Big Dat...
 

Plus de Ontico

Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
Ontico
 

Plus de Ontico (20)

One-cloud — система управления дата-центром в Одноклассниках / Олег Анастасье...
One-cloud — система управления дата-центром в Одноклассниках / Олег Анастасье...One-cloud — система управления дата-центром в Одноклассниках / Олег Анастасье...
One-cloud — система управления дата-центром в Одноклассниках / Олег Анастасье...
 
Масштабируя DNS / Артем Гавриченков (Qrator Labs)
Масштабируя DNS / Артем Гавриченков (Qrator Labs)Масштабируя DNS / Артем Гавриченков (Qrator Labs)
Масштабируя DNS / Артем Гавриченков (Qrator Labs)
 
Создание BigData-платформы для ФГУП Почта России / Андрей Бащенко (Luxoft)
Создание BigData-платформы для ФГУП Почта России / Андрей Бащенко (Luxoft)Создание BigData-платформы для ФГУП Почта России / Андрей Бащенко (Luxoft)
Создание BigData-платформы для ФГУП Почта России / Андрей Бащенко (Luxoft)
 
Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
Готовим тестовое окружение, или сколько тестовых инстансов вам нужно / Алекса...
 
Новые технологии репликации данных в PostgreSQL / Александр Алексеев (Postgre...
Новые технологии репликации данных в PostgreSQL / Александр Алексеев (Postgre...Новые технологии репликации данных в PostgreSQL / Александр Алексеев (Postgre...
Новые технологии репликации данных в PostgreSQL / Александр Алексеев (Postgre...
 
PostgreSQL Configuration for Humans / Alvaro Hernandez (OnGres)
PostgreSQL Configuration for Humans / Alvaro Hernandez (OnGres)PostgreSQL Configuration for Humans / Alvaro Hernandez (OnGres)
PostgreSQL Configuration for Humans / Alvaro Hernandez (OnGres)
 
Inexpensive Datamasking for MySQL with ProxySQL — Data Anonymization for Deve...
Inexpensive Datamasking for MySQL with ProxySQL — Data Anonymization for Deve...Inexpensive Datamasking for MySQL with ProxySQL — Data Anonymization for Deve...
Inexpensive Datamasking for MySQL with ProxySQL — Data Anonymization for Deve...
 
Опыт разработки модуля межсетевого экранирования для MySQL / Олег Брославский...
Опыт разработки модуля межсетевого экранирования для MySQL / Олег Брославский...Опыт разработки модуля межсетевого экранирования для MySQL / Олег Брославский...
Опыт разработки модуля межсетевого экранирования для MySQL / Олег Брославский...
 
ProxySQL Use Case Scenarios / Alkin Tezuysal (Percona)
ProxySQL Use Case Scenarios / Alkin Tezuysal (Percona)ProxySQL Use Case Scenarios / Alkin Tezuysal (Percona)
ProxySQL Use Case Scenarios / Alkin Tezuysal (Percona)
 
MySQL Replication — Advanced Features / Петр Зайцев (Percona)
MySQL Replication — Advanced Features / Петр Зайцев (Percona)MySQL Replication — Advanced Features / Петр Зайцев (Percona)
MySQL Replication — Advanced Features / Петр Зайцев (Percona)
 
Внутренний open-source. Как разрабатывать мобильное приложение большим количе...
Внутренний open-source. Как разрабатывать мобильное приложение большим количе...Внутренний open-source. Как разрабатывать мобильное приложение большим количе...
Внутренний open-source. Как разрабатывать мобильное приложение большим количе...
 
Подробно о том, как Causal Consistency реализовано в MongoDB / Михаил Тюленев...
Подробно о том, как Causal Consistency реализовано в MongoDB / Михаил Тюленев...Подробно о том, как Causal Consistency реализовано в MongoDB / Михаил Тюленев...
Подробно о том, как Causal Consistency реализовано в MongoDB / Михаил Тюленев...
 
Балансировка на скорости проводов. Без ASIC, без ограничений. Решения NFWare ...
Балансировка на скорости проводов. Без ASIC, без ограничений. Решения NFWare ...Балансировка на скорости проводов. Без ASIC, без ограничений. Решения NFWare ...
Балансировка на скорости проводов. Без ASIC, без ограничений. Решения NFWare ...
 
Перехват трафика — мифы и реальность / Евгений Усков (Qrator Labs)
Перехват трафика — мифы и реальность / Евгений Усков (Qrator Labs)Перехват трафика — мифы и реальность / Евгений Усков (Qrator Labs)
Перехват трафика — мифы и реальность / Евгений Усков (Qrator Labs)
 
И тогда наверняка вдруг запляшут облака! / Алексей Сушков (ПЕТЕР-СЕРВИС)
И тогда наверняка вдруг запляшут облака! / Алексей Сушков (ПЕТЕР-СЕРВИС)И тогда наверняка вдруг запляшут облака! / Алексей Сушков (ПЕТЕР-СЕРВИС)
И тогда наверняка вдруг запляшут облака! / Алексей Сушков (ПЕТЕР-СЕРВИС)
 
Как мы заставили Druid работать в Одноклассниках / Юрий Невиницин (OK.RU)
Как мы заставили Druid работать в Одноклассниках / Юрий Невиницин (OK.RU)Как мы заставили Druid работать в Одноклассниках / Юрий Невиницин (OK.RU)
Как мы заставили Druid работать в Одноклассниках / Юрий Невиницин (OK.RU)
 
Разгоняем ASP.NET Core / Илья Вербицкий (WebStoating s.r.o.)
Разгоняем ASP.NET Core / Илья Вербицкий (WebStoating s.r.o.)Разгоняем ASP.NET Core / Илья Вербицкий (WebStoating s.r.o.)
Разгоняем ASP.NET Core / Илья Вербицкий (WebStoating s.r.o.)
 
100500 способов кэширования в Oracle Database или как достичь максимальной ск...
100500 способов кэширования в Oracle Database или как достичь максимальной ск...100500 способов кэширования в Oracle Database или как достичь максимальной ск...
100500 способов кэширования в Oracle Database или как достичь максимальной ск...
 
Apache Ignite Persistence: зачем Persistence для In-Memory, и как он работает...
Apache Ignite Persistence: зачем Persistence для In-Memory, и как он работает...Apache Ignite Persistence: зачем Persistence для In-Memory, и как он работает...
Apache Ignite Persistence: зачем Persistence для In-Memory, и как он работает...
 
Механизмы мониторинга баз данных: взгляд изнутри / Дмитрий Еманов (Firebird P...
Механизмы мониторинга баз данных: взгляд изнутри / Дмитрий Еманов (Firebird P...Механизмы мониторинга баз данных: взгляд изнутри / Дмитрий Еманов (Firebird P...
Механизмы мониторинга баз данных: взгляд изнутри / Дмитрий Еманов (Firebird P...
 

Dernier

notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.ppt
MsecMca
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
Epec Engineered Technologies
 
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
HenryBriggs2
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
ssuser89054b
 
DeepFakes presentation : brief idea of DeepFakes
DeepFakes presentation : brief idea of DeepFakesDeepFakes presentation : brief idea of DeepFakes
DeepFakes presentation : brief idea of DeepFakes
MayuraD1
 
"Lesotho Leaps Forward: A Chronicle of Transformative Developments"
"Lesotho Leaps Forward: A Chronicle of Transformative Developments""Lesotho Leaps Forward: A Chronicle of Transformative Developments"
"Lesotho Leaps Forward: A Chronicle of Transformative Developments"
mphochane1998
 

Dernier (20)

HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptxHOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
 
Block diagram reduction techniques in control systems.ppt
Block diagram reduction techniques in control systems.pptBlock diagram reduction techniques in control systems.ppt
Block diagram reduction techniques in control systems.ppt
 
notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.ppt
 
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
 
Employee leave management system project.
Employee leave management system project.Employee leave management system project.
Employee leave management system project.
 
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
 
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptxS1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
 
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
COST-EFFETIVE  and Energy Efficient BUILDINGS ptxCOST-EFFETIVE  and Energy Efficient BUILDINGS ptx
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
 
Learn the concepts of Thermodynamics on Magic Marks
Learn the concepts of Thermodynamics on Magic MarksLearn the concepts of Thermodynamics on Magic Marks
Learn the concepts of Thermodynamics on Magic Marks
 
Unleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leapUnleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leap
 
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
scipt v1.pptxcxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 
Computer Lecture 01.pptxIntroduction to Computers
Computer Lecture 01.pptxIntroduction to ComputersComputer Lecture 01.pptxIntroduction to Computers
Computer Lecture 01.pptxIntroduction to Computers
 
AIRCANVAS[1].pdf mini project for btech students
AIRCANVAS[1].pdf mini project for btech studentsAIRCANVAS[1].pdf mini project for btech students
AIRCANVAS[1].pdf mini project for btech students
 
DeepFakes presentation : brief idea of DeepFakes
DeepFakes presentation : brief idea of DeepFakesDeepFakes presentation : brief idea of DeepFakes
DeepFakes presentation : brief idea of DeepFakes
 
School management system project Report.pdf
School management system project Report.pdfSchool management system project Report.pdf
School management system project Report.pdf
 
"Lesotho Leaps Forward: A Chronicle of Transformative Developments"
"Lesotho Leaps Forward: A Chronicle of Transformative Developments""Lesotho Leaps Forward: A Chronicle of Transformative Developments"
"Lesotho Leaps Forward: A Chronicle of Transformative Developments"
 
Work-Permit-Receiver-in-Saudi-Aramco.pptx
Work-Permit-Receiver-in-Saudi-Aramco.pptxWork-Permit-Receiver-in-Saudi-Aramco.pptx
Work-Permit-Receiver-in-Saudi-Aramco.pptx
 
data_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfdata_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdf
 

Не так страшен терабит / Вячеслав Ольховченков (Integros)

  • 2. Profile • Full stack video solution • Adaptive multibitrate streaming • HLS, DASH & mp4 support, up to 1080p, native playback! • Advertising standards supported: IMA, VPAID, VAST 2,3 • User platforms: desktop, mobile, STB, TV • Support embedding into application • Regions: CIS, EU, US, South America • Content: UGC, advertising, professional 2
  • 3. Load profile • Peak bandwidth 1Tbit/s (200Gbit/s @ 2015) • Peak bandwidth per edge 74Gbit/s (40Gbit/s @2015) • Daily traffic 6+PTB • Views per day 40M • Aggregated storage cap: 1.3PB (0.7PB used) 3
  • 4. What’s new? • Moved to html5 player • Dropped Flash & HDS support • SSL support • P2P in production 4
  • 6. Edge (features) • Cost-effectively, high-performance • Store and delivery content to user • Check URL signature • Collect per-client stats • Collect per-stream stats • 60+K simulations users • 74Gb/s output 6
  • 7. Edge (architecture) • FreeBSD 11 ZFS ZMIRROR 32TB HDD ZFS L2ARC 3.5TB SSD ZFS ARC 190GB RAM Nginx, Lua NIC 2x40Gbit Internet 7
  • 8. Edges history • V1 (20Gbit), 2012 • V2 (42Gbit), 2013 • V3 (52Gbit), 2015 • V4 (74Gbit), 2016 8
  • 9. Edge v4.0 (configuration) • Dual E5-2650 v4 @ 2.20GHz • 128GB RAM • 8 SSD 500GB • 18 HDD 4+TB • DualPort Chelsio T580-LP-CR Fail: 54.55Gbit/s limited (CPU idle: 28%) 9
  • 10. Edge PCIe limitation PCIe Speed Encoding Total (8x) v1.0 2.5 Gb/s 8B/10B 16 Gbit/s v2.0 5.0 Gb/s 8B/10B 32 Gbit/s v3.0 8.0 Gb/s 128B/130B 63 Gbit/s Gen PHY DLL ECRC (opt) TLP header Total Data 128 Data 256 1/2 1+1 2+4 4 16 24/28 .84/.82 .91/.90 3.0 4 2+4 4 16 26/30 .83/.81 .91/.90 TLP overhead (64bit): Total: 8*8/130*128*256/(256+30) = 56.4 10
  • 11. Edge v4.1 (configuration) • Dual E5-2650 v4 @ 2.20GHz • 256GB RAM • 10 SSD 500GB • 18 HDD 4+TB • DualPort Chelsio T62100-LP-CR / Solarflare SFC9200 • Solarflare: firmware QSFP+ issues, BIOS bugs • Chelsio: firmware uneven distributed, link down 11
  • 12. Edge load profile 74 Gbit/s 12
  • 13. Not all SSD are equally useful • Crucial MX100 (ok) • Crucial BX100 (ok) • Crucial MX200 (medium speed ok) • Crucial BX200 • Crucial MX300 • Fail at high speed (about ~35Gbit). 13
  • 14. SSD comparison MX100 MX200 BX200 MX300 850Pro 850Evo read 475 482 463 456 506 495 write 443 335 232 310 457 391 mixed 423 326 248 318 478 398 read4k 24.8 24.4 25.2 21.6 32.4 38 write4k 83.3 68 43.9 60.2 57.6 77.9 mixed4k 24.6 19.9 17 17.8 34.7 35 -15% -33% -23% +17% +12% 14
  • 15. SSL: performance drop • Expectations (from Google): no performance drop. • Reality: performance drop, load rise about 1% CPU load per Gigabit (0.72GHz/Gbit => 1.22GHz/Gbit) • Measurements (per Gigabit): • SSL off: 1.11% sys + 0.25% user • SSL on: 0.75% sys + 1.55% user • 0.72 busy @ 33Gbit (Dual E5-2650v4 @ 2.20GHz), expected: 45Gbit 15
  • 16. P2P (WebRTC) • Goals: • cost reduce • QoS for distant sites • Streamroot • Hot content • Up to 70% offload during spikes • 90% for top 50 streams 16
  • 17. P2P Architecture 1. Initialization of the video 2. Connection to the tracker 3. Direct exchanges with the network of viewers 1. 2. 3. Pool of available viewers 17
  • 18. P2P Architecture • Viewer authenticates and initializes the connection with Integros load balancer • At the same time, viewer connects to Streamroot tracker and requests available peers • Tracker sends back intelligently-selected list of peers that is updated throughout the session according to bitrate and other criteria • A P2P WebRTC DataChannel is created directly between viewers to exchange segments 18
  • 19. P2P wins • Server offload • Up to 70% all week • 277Gbps offload on peak, 241 on 95/5 !! • Live streams • Start at 23 concurent viewers (50+% offloads) 19
  • 20. Metric P2P ON P2P OFF Difference Buffering rate 0.84% 1.05% -20% Average Bitrate 1.08 Mbps 1.06 Mbps +1.9% Session duration avg 1088s 1065s +2.2% P2P wins (QoS) • Doesn’t add ANY latency, for VoD or Live! • The user experience is EXACTLY the same 20
  • 21. P2P Issues and Current Status • Device and browser support (80-90%% devices supported) • Desktop • Chrome, FF, Opera Yandex OK • Safari 11+ OK since September !! • Edge: OK • Mobile • Webview • Chrome Android Ok • Safari iOS11: soon • Apps • Android SDK OK (Exoplayer, Nexplayer, etc) • iOS SDK OK (AVPlayer & Webview) • STBs & SmartTVs: • Android TV OK • Apple TV soon • Smart TV: not yet  21
  • 22. P2P Future • Compatibility growth • STBs • Smart TVs! • Continue R&D • Better QoS algos • Better P2P algos 22

Notes de l'éditeur

  1. 1. Viewer authenticates and initializes the connection with Integros load balancer 2. At the same time, viewer connects to Streamroot tracker and requests available peers 3. Tracker sends back intelligently-selected list of peers that is updated throughout the session according to bitrate and other criteria 4. A P2P WebRTC DataChannel is created directly between viewers to exchange segments