Video Cloud Computing White Paper
Video Cloud Computing White Paper
Video Cloud Computing White Paper
>> www.thomson-networks.com
WHITE PAPER
1.
OVERVIEW .................................................................................................................................................... 3
1.1
1.2
DEFINITIONS .............................................................................................................................................. 3
VIDEO CLOUD ............................................................................................................................................ 5
2.
3.
4.
5.
CONCERNS .................................................................................................................................................... 9
CONCLUSION ...................................................................................................................................................... 10
WHITE PAPER
1. OVERVIEW
Nowadays, everyone seems to be talking about cloud computing.
But what does cloud computing mean for video delivery and especially for live processing?
1.1 Definitions
Cloud computing is a way of making computing resources quickly available for any computing task. Cloud
computing assumes the availability of an IP network for connecting all the various computing nodes.
This model arose from observations of IT deployment within companies. Company departments use specific
software dedicated to their business areas general accounting software, customer relationship
management software or a web portal dedicated to promotion over the Internet and this software is
usually installed on separate computers. In most cases, even if this software is running continuously it is not
using 100% of the CPU resources 24/7.
Cloud computing allows deployment of this software on a farm of standard servers or even a single
computer, and it runs as if it were located on standalone computers. IT management is simplified, usage of
computing power is optimized and less hardware is required for the same job.
A cloud architecture has several deployment models. In a private cloud architecture, the computer farm is
located in an internal data center, while in a public cloud use case, the computers are available via the
Internet and rented from a third-party company.
Public cloud providers include many types of services in their offerings. In addition to rental of computing
resources they can guarantee architecture availability. In the event of a hardware component failure, the
software is automatically restarted on another server.
Public cloud providers also offer a library of software for rent on their system. This library contains standard
or third-party applications.
The terminology differs depending on the service provided.
Infrastructure as a Service (IaaS) consists of the rental of hardware computer resources. Some low-level
services such as redundancy management could be included. The customer is required to provide the
complete software stack. IaaS means renting just the computers without any software installed on them.
The Platform as a Service (PaaS) model offers low-level software modules such as the operating system in
addition to IaaS. The customer provides the application to be run on the platform. Both IaaS and PaaS require
development skills for implementing and maintaining the chosen solutions.
Finally, in the Software as a Service (SaaS) use case, in addition to the platform the end software running on it
is also rented.
3
WHITE PAPER
Video file transcoding is a typical SaaS application. Companies are now offering the option to upload video
files via the Internet and their web portal, and in return for payment customers get transcoded versions in
numerous formats and bitrates.
In a SaaS setting, the hardware dimension disappears and even the cloud provider may also become invisible.
Customers just deal with a service provider through a website.
remote
servers
WHITE PAPER
Cloud
processing
IP network
(Internet or
dedicated lines)
reception
modulator
management system
In the same manner, data resulting from processing for example, a TS multiplex with compressed video and
audio needs to be downloaded from the cloud to be modulated before delivery via a satellite network.
Some tasks such as acquisition and modulation need to remain in the video delivery platform and performed
by locally operated devices. In between those tasks, any type of processing can be performed in the cloud.
Video and audio transcoding, logo insertion, splicing, ad insertion, OTT packaging, scrambling, and statistical
multiplexing are all cloud-compatible.
Because cloud computers are connected to the Internet or are provided by worldwide cloud providers, global
data transfer could also be simplified by a cloud architecture.
With an OTT application, for example, data can be sent directly to a content delivery network (CDN).
WHITE PAPER
video delivery
to cloud
processing
Cloud
processing
IP network
(Internet or
dedicated lines)
baseband
sources
to Internet
delivery
CDN
encoders
(SDI-to-IP gateways)
compressed
sources over IP
management system
WHITE PAPER
Application A
AComputernstan
OS
ceA
Application B
AComputernstan
OS
ceB
Application C
AComputernstan
OS
ceC
Application D
AComputernstan
OS
ceD
Application E
AComputernstan
OS
ceE
Hardware A
(Server 1)
Hardware B
(Server 1)
Hardware C
(Server 1)
Hardware D
(Server 1)
Hardware E
(Server 1)
Hypervisor
Manager
VM
VM
VM
VM
VM
Application A
AComputernstan
OS
ceA
Application B
AComputernstan
OS
ceB
Application C
AComputernstan
OS
ceC
Application D
AComputernstan
OS
ceD
Application E
AComputernstan
OS
ceE
OS
Hypervisor
Hypervisor
Hardware
Hardware (Server)
Hardware (Server)
OpenStack is a hypervisor manager that is relatively well-known and deployed by cloud providers.
Amazon EC2 runs a proprietary hypervisor and proprietary manager which allow implementation of several
business models.
More than 10 different virtual computer models are available in the Amazon offering, including GPU-based
VMs for task-intensive processing. Virtual computers can be booked in 5 different areas all over the globe. An
Amazon customer can even resell their virtual computer capabilities via auction-based websites such as eBay.
WHITE PAPER
WHITE PAPER
5. CONCERNS
There are always two sides of a coin and cloud computing is no exception.
Cloud computing becomes expensive when used 365 days a year compared to a physical video delivery
solution where the hardware is bought and not rented.
Availability could also be an issue since it has a strong impact on broadcasters. NetFlix suffered a 3-day
blackout in 2008 and Amazon EC2 a failure lasting several hours in 2012. Current public cloud solutions offer
99.9% uptime at their best, i.e. 2 hours of downtime per quarter for a 24-hour encoding service.
Private cloud solutions or cloud architectures can offer a better QoS but at a higher cost.
While cloud providers guarantee availability, the risk can be considered too high by some broadcasters.
Operating cloud as a PaaS or IaaS can be quite complex. Customers building their own solution may have to
coordinate several stakeholders such as the Telco operator for the cloud connection, the cloud provider and
the software provider. Establishing responsibility in the event of failure may not always be easy.
SaaS is the optimal solution as a single company is responsible for selecting the cloud provider and deploying
the software in the cloud. But Saas solutions are quite different from the currently deployed architectures.
Having a Saas provider managing both local video platform resources and cloud resources ensures a smoother
transition between the two environments.
Depending on the content value, some customers may also have concerns with outsourcing to a public cloud
company. Scrambling and conditional access might be needed to reduce the risk. For highly sensitive content, a
private data center solution may remain the only acceptable solution.
WHITE PAPER
CONCLUSION
Cloud computing is an additional tool available to broadcasters seeking flexibility.
Currently deployed solutions mainly address file encoding. A good cloud solution for real-time video platform
delivery will need to offer several advantages to seduce broadcasters in terms of content protection,
availability, and flexibility. Under these conditions and in addition to premium processing and compression,
cloud computing may become more widespread among broadcasters.
CONTACT INFORMATION
Please contact your sales representative or
visit our website for all contact details.
http://www.thomson-networks.com
Copyright 2015 Thom son Video Netw orks. All rights reserved. All other trade nam es referenced are service m arks, tradem arks, or registered tr adem arks
of their respective companies. Specifications subject to change w ithout notice. CDT-5226D.
>> www.thomson-networks.com
10