1
Cloud Transition Patterns for Media
Enterprises
Shailendra Mathur
Vice President of Architecture, Avid
2 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Topics
 Understanding current media production infrastructure
 Different “Cloud” definitions evolving in the media Industry
 Lift & Shift methods vs Micro-services based implementations
 Avid Media Central Platform based communication fabric across
hosting models
3 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Source: http://ngcodec.com/fpga-encoder-markets/
Narrowing the problem domain
4 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Existing production workflows and infrastructure
Asset Management & Storage
5 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Machine Rooms/Datacenter
• VM/Bare Metal
• Single Customer Instance
• Network Storage
• Edge servers (SDI/IP)
Interpretation of “Cloud” in the media industry
On-prem Cloud
• VMs & Containers
• Single customer Instance
• Private
• Network/Cloud Storage
Private Cloud
• VMs & Containers
• Single customer
Instance
Public Cloud
• VM & Containers
• Public Cloud Provider
• Multiple customers
“Local” Cloud
• VMs & Containers
• Managed Service
provider
• Multiple customer
CoLo Datacenter
• Storage
• Edge Servers (IP)
Cloud “On-Prem”
WAN
High
Bandwidth
Network Link
Hybrid Cloud
Ali, AWS, MSFT, IBM, Google…
HP, Dell, VMWare, IBM, MSFT, Huawei…
6 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Media services hosting patterns
Host OS
Any
App
Any
App
Any
App
Bare Metal
Hypervisor Layer
Guest
OS
App1
Compute, Network, Storage,
VDI,, A/V IP interfaces
Bins/
Libs
Guest
OS
App2
Bins/
Libs
Guest
OS
App3
Bins/
Libs
CloudProvider
(IaaS)
AppProvider
(VirtualMachines)
Virtualized
Cloud or On Prem
Operating System
App1
Compute, Network, Storage
Bins/
Libs
CloudProvider
(PaaS)
AppProvider
(Containers)
Bins/Libs
App2 App3
Micro-Services Containers
Cloud Native
Docker Engine
AppProvider
(Machines)
Compute, Network, Storage,
Display, A/V IP/SDI
interfaces, Controllers
Web Client
Web
Client
Thick
Client
Web
Client
Thick
Client
7 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Why Micro-Services?
Multi-function software appliance
8 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
Multi-function software appliance
9 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
…vs multiple functional micro-services
1 2 3 4
10 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
Service hosting and communication
App Provider
Communication Fabric
App Provider
Communication Fabric
Cloud Provider
Communication Fabric
Bare Metal Virtualized
Cloud or On Prem
Micro-services Containers
Cloud Native
11 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
Service hosting and communication
Common communication event and message bus
(App Provider)
Bare Metal Virtualized
Cloud or On Prem
Micro-services Containers
Cloud Native
12 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Why a message bus?
1
3
4 2
5
6
7
8
9
10
Point to Point Traditional communication fabric
13 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
Traditional communication fabric
14 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015
…vs pub-sub service interactions
5
Event Enabled Microservices
(Publish/Subscribe)
1 2 3 4
15 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Common service communication framework
Cloud Native
Federated Message BusFederated Message Bus Federated Message| Bus
Bare Metal Virtualized Data Center Cloud
16 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Media ingest & processing example
Federated Message BusFederated Message Bus Federated Message| Bus
Cloud Upload
Bare Metal Virtualized Data Center Cloud
17 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Avid Media Central framework
Service
Management
Service
Execution
User Admin
& Config
Connectivity
Toolkit
Cloud Native
API Upstreams
Federated Message BusFederated Message Bus Federated Message| Bus
18 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Summary
 “Cloud” has different hosting interpretations
 The Media Industry looking to implement both Lift and Shift and
Cloud native models
 Avid Media Central Framework bridges the different hosting models
20 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Multiple message & data model formats
…
…
21 Confidential and Proprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017.
Common message data model interpreters
…
…

Cloud Transition Patterns for Media Enterprises

  • 1.
    1 Cloud Transition Patternsfor Media Enterprises Shailendra Mathur Vice President of Architecture, Avid
  • 2.
    2 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Topics  Understanding current media production infrastructure  Different “Cloud” definitions evolving in the media Industry  Lift & Shift methods vs Micro-services based implementations  Avid Media Central Platform based communication fabric across hosting models
  • 3.
    3 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Source: http://ngcodec.com/fpga-encoder-markets/ Narrowing the problem domain
  • 4.
    4 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Existing production workflows and infrastructure Asset Management & Storage
  • 5.
    5 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Machine Rooms/Datacenter • VM/Bare Metal • Single Customer Instance • Network Storage • Edge servers (SDI/IP) Interpretation of “Cloud” in the media industry On-prem Cloud • VMs & Containers • Single customer Instance • Private • Network/Cloud Storage Private Cloud • VMs & Containers • Single customer Instance Public Cloud • VM & Containers • Public Cloud Provider • Multiple customers “Local” Cloud • VMs & Containers • Managed Service provider • Multiple customer CoLo Datacenter • Storage • Edge Servers (IP) Cloud “On-Prem” WAN High Bandwidth Network Link Hybrid Cloud Ali, AWS, MSFT, IBM, Google… HP, Dell, VMWare, IBM, MSFT, Huawei…
  • 6.
    6 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Media services hosting patterns Host OS Any App Any App Any App Bare Metal Hypervisor Layer Guest OS App1 Compute, Network, Storage, VDI,, A/V IP interfaces Bins/ Libs Guest OS App2 Bins/ Libs Guest OS App3 Bins/ Libs CloudProvider (IaaS) AppProvider (VirtualMachines) Virtualized Cloud or On Prem Operating System App1 Compute, Network, Storage Bins/ Libs CloudProvider (PaaS) AppProvider (Containers) Bins/Libs App2 App3 Micro-Services Containers Cloud Native Docker Engine AppProvider (Machines) Compute, Network, Storage, Display, A/V IP/SDI interfaces, Controllers Web Client Web Client Thick Client Web Client Thick Client
  • 7.
    7 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Why Micro-Services? Multi-function software appliance
  • 8.
    8 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 Multi-function software appliance
  • 9.
    9 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 …vs multiple functional micro-services 1 2 3 4
  • 10.
    10 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 Service hosting and communication App Provider Communication Fabric App Provider Communication Fabric Cloud Provider Communication Fabric Bare Metal Virtualized Cloud or On Prem Micro-services Containers Cloud Native
  • 11.
    11 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 Service hosting and communication Common communication event and message bus (App Provider) Bare Metal Virtualized Cloud or On Prem Micro-services Containers Cloud Native
  • 12.
    12 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Why a message bus? 1 3 4 2 5 6 7 8 9 10 Point to Point Traditional communication fabric
  • 13.
    13 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 Traditional communication fabric
  • 14.
    14 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2015 …vs pub-sub service interactions 5 Event Enabled Microservices (Publish/Subscribe) 1 2 3 4
  • 15.
    15 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Common service communication framework Cloud Native Federated Message BusFederated Message Bus Federated Message| Bus Bare Metal Virtualized Data Center Cloud
  • 16.
    16 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Media ingest & processing example Federated Message BusFederated Message Bus Federated Message| Bus Cloud Upload Bare Metal Virtualized Data Center Cloud
  • 17.
    17 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Avid Media Central framework Service Management Service Execution User Admin & Config Connectivity Toolkit Cloud Native API Upstreams Federated Message BusFederated Message Bus Federated Message| Bus
  • 18.
    18 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Summary  “Cloud” has different hosting interpretations  The Media Industry looking to implement both Lift and Shift and Cloud native models  Avid Media Central Framework bridges the different hosting models
  • 20.
    20 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Multiple message & data model formats … …
  • 21.
    21 Confidential andProprietary Information. Please do not copy, forward, redistribute, or publish. Avid ©2017. Common message data model interpreters … …

Editor's Notes

  • #4 The part of production chain in focus
  • #5 Introduce Production Workflows Ingest processes Workstation & SW and SW alliances – primarily CPU compute, FPGA for encode Audio, Video Ancillary Data high bandwidth SDI connected, now IP Creative Applications Typically Workstations Some Thin Clien Apps usign Web UX and SW appliances CPU, local disk and GPUs (for compute and display) Playout/distribution Workstation & SW and SW alliances – primarily CPU compute, GPU for Gfx playout/branding, Audio, Video Ancillary Data high bandwidth SDI connected, now IP Asset mgmnt Shared Storage (Tiered)– Nearline, Online (real-time), Parking Storage and Archive stirage SW Appliances
  • #6 Different Cloud Hosting models for Media Enterprises Definition of Cloud as per NIST
  • #7 Micro-services architecture Cloud Compatible vs Cloud native VM base dlift and Shift vs Containers Introduce “Cloud Portable” to avoid vendor lock Software appliance model – server based apps New boundaries – Media Client virtualization Evolution of Cloud Microservices Framworks https://www.slideshare.net/biztalk360/distributed-computing-made-easy-with-service-fabric?next_slideshow=1 Azure Fabric AWS Lambda Google Functions
  • #11 Micro-services architecture Cloud Compatible vs Cloud native VM base dlift and Shift vs Containers Introduce “Cloud Portable” to avoid vendor lock Software appliance model – server based apps New boundaries – Media Client virtualization Evolution of Cloud Microservices Framworks https://www.slideshare.net/biztalk360/distributed-computing-made-easy-with-service-fabric?next_slideshow=1 Azure Fabric AWS Lambda Google Functions
  • #12 Micro-services architecture Cloud Compatible vs Cloud native VM base dlift and Shift vs Containers Introduce “Cloud Portable” to avoid vendor lock Software appliance model – server based apps New boundaries – Media Client virtualization Evolution of Cloud Microservices Framworks https://www.slideshare.net/biztalk360/distributed-computing-made-easy-with-service-fabric?next_slideshow=1 Azure Fabric AWS Lambda Google Functions
  • #15 - Cloud for publish subscribe.
  • #16 Solution Thick or Thin Apps Authenticated REST API calls Pub Sub arch Supports Micro services natively Adapter svc to traditional Bare Metal and VM based SW appliances Federated message Bus Static and Dynamic Deployment
  • #17 Solution Thick or Thin Apps Authenticated REST API calls Pub Sub arch Supports Micro services natively Adapter svc to traditional Bare Metal and VM based SW appliances Federated message Bus Static and Dynamic Deployment
  • #18 Solution Thick or Thin Apps Authenticated REST API calls Pub Sub arch Supports Micro services natively Adapter svc to traditional Bare Metal and VM based SW appliances Federated message Bus Static and Dynamic Deployment
  • #22 Share Libs for Message interpretation, Events and Data model