284 23-3171 | Uen

WebRTC: enhancing the web

with real-time communication
capabilities
G ör a n a p E r i k s s on a n d S t e fa n H å k a n s s on

Since the early days of HTML and the Mosaic browser, the web platform has become
a universal interface working across device, operating system, language and cultural barriers.

The web platform gives people becoming mainstream. HTML5 is a in collaboration with IETF to secure the
a unique way of viewing a wide clear example of how change, driven Box b  protocol aspects of native browser sup-
variety of content, it provides by a combination of factors, creates WebRTC port for a real-time communication API
developers with a write-once something new and is itself a catalyst This refers to the is the first step. Early implementations
deploy-everywhere model, and for more innovation. These factors work being car- of the API are available in a prototype
supports service providers in include increasingly widespread fixed ried out within browser from Ericsson and developer
deploying services with global and mobile broadband, consumer and W3C and IETF versions of Google Chrome. So far, the
reach. The waves of browser-tech- enterprise demand for multimedia to standardize results are promising, but the path to
nology enhancement, exemplified communication services, and the need support for real- interoperable implementation in all
by HTML5, and the ongoing work for a platform to reach users – with time capabilities major browsers and all device types is
to add real-time communication many types of devices – at a manage- in browsers, long, and success is not guaranteed.
to the web platform create new able cost. allowing web The objective of the standardization
opportunities for combining The current browser-development applications to work is to create an API that enables a
communication and data – and revolution, with HTML5 at its helm, is send data and web application running on any device
improving the user experience. pushing browser capability closer to media streams to send and receive real-time media and
that of native application environments. between devices data streams in a peer-to-peer fashion
HTML5 and browser development Support for enabling real-time commu- over IP networks between the browsers. Running in a
Technical, sociocultural, local and glob- nication (RTC) on the web – WebRTC (see in a peer-to-peer browser, the API will for the most part
al changes in ICT can happen overnight. Box B) – is emerging. The ongoing stan- fashion. use existing protocols and codecs in
People adopt new solutions instantly, dardization work being carried out by the browser core. The design of the API
with cutting-edge technologies rapidly the World Wide Web Consortium (W3C) leaves the web developer free to imple-
ment functionality for finding and
connecting the participants in a com-
BOX A  Terms and abbreviations munication session. The web platform
provides the support for blending com-
3GPP 3rd Generation Partnership Project MMTel multimedia telephony
munication and data across devices and
API application programming interface NAT Network Address Translation
operating systems, and is a key technol-
CSS Cascading Style Sheets OS operating system
ogy if solutions are to meet the wide
DOM Document Object Model PaaS platform as a service
HD high-definition PSTN public switched telephone network
variety of present and future needs.
HDVC HD Visual Communication QoS quality of service
One of the greatest strengths of open-
HTML HyperText Markup Language RCS Rich Communication Suite browser standardization is that it lever-
HTML5 fifth revision of the HTML standard REST Representational State Transfer ages the work of organizations such as
HTTP Hypertext Transfer Protocol RTC real-time communication W3C and IETF, and other highly-com-
ICE Interactive Connection RTP Real-time Transport Protocol petent browser technology communi-
Establishment SCTP Stream Control Transmission Protocol ties – such as Apple-hosted WebKit open
ICT information and communications SDP Session Description Protocol source project, and the Web Hypertext
technology SIP Session Initiation Protocol Application Technology Working Group
IETF Internet Engineering Task Force UDP User Datagram Protocol (WHATWG). These communities share
IMS IP Multimedia Subsystem VPN virtual private network knowledge actively and use iterative
IP Internet Protocol W3C World Wide Web Consortium solution development to evolve the
LTE Long Term Evolution WHATWG Web Hypertext Application web platform step-by-step according to
MBB mobile broadband Technology Working Group developer and user needs.

E r i c s s o n r e v i e w • 1 2012
 3

Web as a platform for applications

The web was not designed with applica- Figure 1  Basic solution
tion development in mind. Over time,
it has grown into a platform that sup- Web server
ports applications well. Improvements Application
have been driven by developers’ need
to create innovative solutions quickly. Web app download and Presentation
control signaling over
JavaScript, the DOM convention and HTTP or WebSocket Find and connect
some other (often proprietary) features
have paved the way for the modern era
of web-based application development Web server platform
in which new APIs are added to improve Device Device
existing apps and put them on par – ICE signaling to
in terms of functionality – with their manage NAT and
Helper firewall traversal
native counterparts. HTML5, (ICE) server Web
The set of features used for develop- CSS, JS application
ing web applications is growing con- Audio, video and
stantly and is often referred to as the APIs data on RTP/UDP APIs
Open Web Platform. The success of a and SCTP IP backbone
given feature is ultimately determined
Browser or Browser or
by whether or not it gets included in the webview webview
major browsers. HTML5 has also been Access Access
used as a synonym for the advance- network network
ments taking place, although it is really
just one part of the open web.
HTML, the markup language of the
web, represents an application and its to the standards will run on the device side-by-side with ongoing standardiza-
data in a structured way, and allows regardless of the operating system and tion. Anyone can propose new features,
the application to be styled with CSS, hardware architecture. In some cases, but developers who can implement a giv-
and to be controlled with JavaScript. the cost of developing applications for en feature and demonstrate the skills
These technologies – HTML, CSS, and multiple devices can be justified, and in and know-how required to make it work
JavaScript – delivered over the HTTP others it is desirable to target individual in the larger context have an immedi-
infrastructure, are the cornerstones device platforms. In most cases, howev- ate advantage in deciding which new
of the Open Web Platform. Combined er, targeting all possible platforms is not functions get accepted and ultimately
with scripting APIs, they form the basis feasible from a cost or time-to-market get implemented in the major brows-
of what makes the web so attractive to perspective. The success of an applica- ers. Implementation work in the
application developers – the way user tion is determined by reach – a solution open-source projects and web standard-
interfaces are described and styled that supports reach at a reasonable cost ization, in particular, is in most cases a
declaratively, and the versatility afford- will consequently be the winner in the meritocracy favoring individual achieve-
ed by a dynamic scripting language. platform competition. ments over design-by-committee.
The proliferation of web applications The W3C plays an important role in
can be partially attributed to the low Browsers and standardization web standardization. However, the con-
entry barrier – snippets of code are cop- There are currently four browser sortium’s specifications are subject to
ied, pasted and shared among devel- engines that are capable of running the business decisions of the various
opers and the only tools needed are a advanced web applications. Two are browser vendors – who in reality deter-
browser and a text editor. More com- open-source: the Mozilla Gecko engine mine which features get implemented.
prehensive sets of utilities are often col- which is used in Firefox; and WebKit, WHATWG is often seen as a browser-
lected in JavaScript libraries, generating which is used in Apple’s Safari, Google’s vendor consortium, but is really a loose
entire frameworks that have their own Chrome, and on several mobile devices amalgamation of individuals working
developer communities, books and even running various OSs and browsers. The in related industries and for the major
conferences in some cases. remaining two engines, Opera’s Presto browser vendors – with the exception
In addition to their appeal to develop- and Microsoft’s Trident (which is used of Microsoft.
ers, the fact that browsers are supported in Internet Explorer) are closed-source.
by almost all devices may ultimately tip Web standardization is current- The road to WebRTC
the scales in favor of using the web, rath- ly progressing in line with develop- In 2006, the first API to facilitate peer-
er than other technologies, for applica- ment in browser-engine technology to-peer connections for web applica-
tion development. If a browser on any – this is particularly noticeable when it tions was added to the WHATWG Web
given device is sufficiently standards- comes to open-source software, where Applications 1.0. This API was developed
compliant, all applications that adhere new features are prototyped in public, purely for sending data.

E r i c s s o n r e v i e w • 1 2012
Putting real-time communication in your pocket
4

it, or sending it to a remote peer. A

Figure 2  A sample vertical service on the web platform LocalMediaStream represents a media
stream from a local camera, micro-
phone, or other media-capture device.
State your symptoms
Ericsson Name: To obtain local streams, web applica-
medical Gender: tions must request access from the
Age:
Pain Allergy
group user (through the getUserMedia func-
Date of birth:
tion). The application specifies the type
Weakness Confusion
of media – audio or video – to which
Social security number:
it requires access. The device selector
Home phone number: Nausea Fever in the browser interface serves as the
Work phone number: mechanism for granting or denying
Mobile phone number: Breathing
Sickness access.
problems
Doctor The PeerConnection represents an
C. Johnson Video Call association to a remote peer – which is
Geriatrics
usually another instance of the same
application running at the remote end.
Once a peer connection is established,
Authorization and request to disclose MediaStream objects can be sent to
my health information
Symptoms Notes Calendar the remote peer. The peer-connection
mechanism uses ICE to traverse NAT
boxes and firewalls. Using ICE pro-
vides a security measure, as it prevents
untrusted web pages and applications
from sending data to hosts that are not
expecting to receive it. Media-plane sig-
naling is carried out-of-band between
In late 2009, a device element and In March 2011, the WHATWG propos- the peers. The API will send signaling
a MediaStream API were added to the al was revamped completely. The device messages that most applications will
WHATWG specification. The new ele- element was discarded in favor of a pure treat as opaque blobs, but which must
ment enabled user-granted access to JavaScript API. This version of the API be transferred securely and efficient-
certain privacy and security-sensitive became detailed and mature enough to ly to the other peer by the web appli-
devices – such as webcams and micro- be implemented. Signaling was updated cation. Each signaling message is fed
phones – and the API was added to han- to be compatible with SDP offer-answer, into the receiving peer connection upon
dle media streams from such devices. including out-of-band transfer of the arrival, and RTP carries the media data.
The API renders content from media- actual SDPs between peers, leaving the The peer-connection concept includes
capture devices stored locally in media application developer to determine how a function that allows web applications
elements, and records them into binary signaling messages should be transport- to send generic data to the remote peer.
blobs for subsequent upload. ed via a server to the remote peer. The main motivation for developing
During the first half of 2010, ini- Decisions regarding how to transport a media control-plane API only is that
tial proposals for a peer-to-peer com- all higher-level signaling not directly many existing services that would ben-
munication API were added to the related to the media plane, such as find- efit from embracing this technology
WHATWG specification. These pro- and-connect mechanisms, were also left already have an established system for
posals used ICE for connection estab- entirely up to the application developer. managing the way users find one anoth-
lishment and RTP over UDP to transfer In August 2011 a W3C working group er. In some cases, a bidirectional, low-
local media streams to a remote peer for forked the WHATWG specification. The latency channel is readily available to
real-time rendering. In addition to the fork then went through a series of revi- carry the signaling messages between
API, the proposals included means for sions based on the ongoing work in the peers. Consider a social-networking ser-
peer-to-peer data communication and WebRTC Working Group, but the basic vice, for example, where users are a fun-
file transfer. Later in the year, Google structure and principles of the original damental part of the service; the way
hosted a workshop involving several specification have remained intact. they find each other and how they inter-
players from the web and telecommu- There are two fundamental concepts act makes up most of the service. Groups
nications industries, including Ericsson. that are part of the API: MediaStream of users form naturally, based on how
During the workshop, it was agreed and PeerConnection. people classify each other as friends, col-
that these companies should continue A MediaStream is an abstract repre- leagues and schoolmates, and services
working publicly in the relevant stan- sentation of an actual media stream. It such as video chat are an obvious exten-
dardization forums rather than form- serves as a handle for managing actions sion to the service. This kind of setting
ing a consortium dedicated to this new on media streams – such as display- creates a need to find a solution for add-
functionality. ing the stream’s contents, recording ing more services to existing platforms

E r i c s s o n r e v i e w • 1 2012
 5

in an unobtrusive manner. The API is powerful smartphones, tablets and technologies to combine services will
such a solution as it only deals with how laptops are reaching appealing price come into focus as the web platform
media streams between peers are set levels – putting devices supporting evolves toward real-time communica-
up, and not with how, when, and why WebRTC-enabled applications into the tion and services. In-process commu-
the peers have decided to communicate. hands of billions of users [1]; nication solutions will become more
When groups of users communicate the capabilities and availability of fixed important – where communication
in a relaxed conference setting on a and mobile broadband are increasing – is an integral part of the application
social networking platform, the peer- while wide-area coverage of the order of designed for a specific process – such
to-peer model may have to be comple- Mbps cannot be expected universally, as social-­networking services, confer-
mented with network infrastructure for the need for transport in densely encing with collaboration support, edu-
connecting more than two or three par- populated areas will trigger a service cation, customer care and health-care
ticipants. The amount of uplink band- evolution using RTC, particularly video; support. Future innovation runs along
width available to the average user may increased urbanization –creating two axes:
be insufficient to communicate with challenges on a global and regional level improving the efficiency of processes
several people simultaneously – but the as well as new business opportunities that use embedded real-time
model could be evolved so that one peer that require innovative solutions and communication; and
becomes a conference mixer. services; creating new services for people with
The application developer and web- increased digitalization of enterprise increasingly digitalized lives who have
service provider determine how to find communication – driven by demands ever more powerful devices.
and connect the peer instances of a web for efficient, and emission-reducing sub- To illustrate the opportunities of an
application. The API requires the brows- stitutes for face-to-face meetings; and RTC-enabled web platform, Ericsson
er core to provide the functionality to the self-perpetuating nature of innova- Research has developed a number of
establish the necessary audio and video tion – which thrives on the availability of application prototypes, including an
channels. The ongoing standardization ­efficient tools for communication. e-health application and a web applica-
work has not yet resulted in a decision Many of the enhancements to commu- tion to access PSTN and MMTel services.
regarding which codec will be used for nication services will include compo-
this purpose. The current assumption nents to support the user in focused Case: talking to your doctor
is that all media and data streams will communication – video conversations Consider a scenario where a person
always be encrypted. and collaboration seldom take place who is feeling unwell contacts a health-
In its most basic implementation, while walking. care support line – whether it is private,
with the devices located in IP networks, Many applications will be RTC- government-provided or even an inter-
the only server infrastructure compo- enabled and the ability of HTML national consortium. To contact the
nents required are ICE servers for tra-
versing NAT boxes, a web server hosting
the web applications, and a solution for Figure 3  Using selected IMS technologies
finding and connecting the two web
application instances.
Web server
Future communication services
The WebRTC API work focuses on pro- Application
viding a set of basic capabilities for web
applications (media handling, connec- Web app download and
tion establishment and NAT traversal) HTTP or WebSocket
leaving the developer free to design the
user experience, combine with web ser- Carrier PaaS services
vices and provide the find-and-­connect Device REST or SIP Device
mechanism (call control). This approach over HTTP or Find and connect
– combined with the inherent capabili- WebSocket (IMS)
ties of the browser to mash cloud and
Web Web
web services to create customized cli- application
ICE Media services
application
ents across devices – results in a power-
ful toolbox for developing innovative
APIs IP APIs
web applications. backbone
For the past two decades, mobile tele-
Browser or Browser or
phony has dominated the evolution of Access
webview webview
real-time, synchronous communica- network Access
tion. This is changing. The direction of
network
future communication services is being
steered by several factors, including:

E r i c s s o n r e v i e w • 1 2012
Putting real-time communication in your pocket
6

case, the patient may have a best-effort

Figure 4  Extending the reach of telecom services personal data plan. So to support the
session, another actor steps in to pro-
vide dynamic QoS by using LTE QoS, for
The internet Telecom network example, offered by the carrier.
Apart from transport, a carrier could
MMTel
also provide ICE servers to manage NAT
Gateway device
box traversal, media servers for trans-
coding, multi-party support, and push
HTTP to
IMS gateway
IMS servers. A carrier could – and perhaps
should – offer the transport needed to
Device Helper connect the participants, using its own
(ICE PSTN transport infrastructure, or that of oth-
server) device er carriers. Depending on the case, and
RCS web assuming the security mechanisms
application ­chosen allow for it, a carrier can pro-
vide value-added media services such
APIs as media recording, speech-to-text
translation and location-based find-
Browser or IP and-­connect, along with sophisticated
webview Access backbone schemes for billing and charging.
network The legal obligations of the service
provider need to be very well defined in
such a scenario. In fact, the legal aspect
of working with the medical sector may
well be the one factor limiting the scope
support line, the person uses a web can also choose to upgrade the session of applications. This example does,
application that initially prompts them to include video. High-quality video, however, highlight the importance
to enter information about the symp- together with data about the patient of the regulatory measures that may
toms they are experiencing. The person and the ability to interact in real time, be needed to support emerging ICT
then triggers a voice call – from within will increase the chance of a successful businesses in e-health and other areas,
the application – which is then assigned course of action being recommended to such as e-education.
to the most appropriate health-care pro- the patient. Blending data and communication
fessional. The health-care professional The system described here must man- in this manner can be applied to simi-
(in this case a nurse) is presented with age data securely. Patient-data-privacy lar situations in which people need to
the symptom information entered by laws must be upheld and access to this contact their service providers or need
the caller, and patient data that the web type of sensitive data is restricted, as to collaborate, such as in an educational
application has retrieved from a medi- is the location and availability of the or remote-working context.
cal database. specialists in the system. The point of
The nurse talks with the patient, veri- this case is to illustrate the potentially Telecom assets in the web
fying and complementing the symptom enormous enterprise value for all par- Telecom networks are an essential part
information to determine the best next ties involved. of developing solutions for future com-
step. In this scenario, the nurse decides A large hospital could implement munication services. Fixed and mobile
to consult a specialist. The application and manage such a solution indepen- broadband access with QoS is a key
presents the nurse with a number of dently. Other organizations with less enabler of evolving web services.
choices, based on information about in-house competence could outsource The arrival of WebRTC on the scene
the patient, including medical history the infrastructure to a carrier or IT ven- has created a new environment in
data. The nurse then selects a specialist dor. Either way, the transport needed which other telecom assets could be
to contact. A customized find-and-con- to provide a good experience is a poten- used to complement the connectivity
nect mechanism using a recommenda- tial challenge for any service provider provided by carriers. The web-based,
tion engine will trigger a notification to – a high-­quality video meeting held via real-time solutions that are required
the specialist through a dedicated appli- a tablet-sized device typically requires are similar to solutions that are already
cation on the device that the specialist speeds of about 800kbps. The transport available in the telecom world. For
has chosen to be able to receive such capacity to support this is not univer- example, functionality for process-
notifications. sal, and service providers would need to ing media – such as transcoding and
The specialist may then choose to join carefully plan for coverage in collabora- multi-party services, call control, push
the ongoing consultation on the notifi- tion with a carrier. There is an opportu- and billing – are candidates for reuse in
cation device or through another one. To nity to offer different mobile broadband the context of many RTC-enabled web
improve the consultation, the specialist QoS levels in real time. In the specific applications.

E r i c s s o n r e v i e w • 1 2012
 7

Some telecom technologies are used in a over MBB will be an attractive proposi-
Göran AP Eriksson
web-platform context and may require tion as part of VPN and mobility offer-
additional functionality. Certain chang- ings to enterprises. is currently an expert in
es are necessary, such as the introduc- Communication Services
tion of an HTTP-to-SIP gateway between Conclusion Architecture at Ericsson
the device and the network, basic ICE Over the past two decades, internet and Research Multimedia
support in the 3GPP IMS session border web technologies have played a central Technologies. Since
gateway, and support for WebSocket. role in communications. With the emer- joining Ericsson in 1991, he has worked
Generic web-­platform capabilities need gence of HTML5 and other standards in many different fields, including GSM,
to be understood and implemented in such as WebRTC, this role is becoming GPRS, IMS, IETF and W3C technologies.
He holds an M.Sc. in physics from the
the event of such changes. This is par- even more significant.
Royal Institute of Technology (KTH) in
ticularly important when it comes to In the short term, an RTC-enabled
Stockholm, Sweden.
security mechanisms, whether for web platform will be an essential com-
cross-domain communication or for plement to native applications. In
providing access to a user’s camera. the long term, it has the potential to Stefan Håkansson
Adding RTC to the web platform compete successfully with such envi-
is currently working as a
creates new requirements such as how ronments in all but a few application
manager at Ericsson
to authenticate the identity of the peer, segments.
Research Multimedia
or how to harmonize the web platform The inherent capability of the web
Technologies. Since joining
and mobile broadband solutions for QoS platform to provide a cross-device appli-
Ericsson in 1985, he has worked in many
reasons. cation environment with combined ser-
different fields, including Airborne Radar
Using the telecom infrastructure to vices makes it a leading integration tool.
systems, Radio Base Station systems
build a web application goes beyond Complemented with real-time capabil- and Multimedia related research. He
the use of one individual technology. By ities, the web will open up the poten- holds an M.Sc. in electrical engineering
leveraging existing telecom interwork- tial to create multiple new services and from the Chalmers University of
ing relationships, carriers can provide applications where context-specific data Technology in Gothenburg, Sweden.
transport and call control beyond their is used in combination with embedded
own reach. This type of federation mod- communication functions. This will put
el will probably evolve organically – ini- the web in a good position to become the
Acknowledgements
tially involving just a few players and platform of choice for multiple services.
Per-Erik Brodin, Adam Bergkvist,
expanding over time but leveraging on The WebRTC API promises to pro-
Vladimir Katardjiev
the existing and evolving solutions for vide good tools for web developers, but
interworking between communication though this framework adds to the capa-
services, such as RCS and HDVC. bilities of the web platform, there are
In addition to using telecom technol- still pieces missing.
ogies to provide enablers in browser-to- To support applications using an
browser communication, another set RTC-enabled web platform, carriers
of solutions involves using the web plat- could provide transport services bun-
form to supply clients to telecom com- dled together with higher-layer services.
munication services, such as PSTN and Gateways for interfacing with existing
MMTel services. In this case, a gateway and emerging telecom network servic-
to manage the differences between the es, such as PSTN and MMTel, will provide
networks and the basic technologies additional value for service providers.
would be required. By making the most of existing inter-
Carriers could offer the services working agreements with other carriers
described here to providers as a bundle, to provide global reach, such a portfolio
often including transport. For service could have a positive impact on enter-
providers to cope with the expect- prises that use communication as part
ed increase in ICT solutions that have of their business processes, creating
both local and global reach, bundles new business opportunities and increas- References
that include transport over the access ing revenue.
network and backbone in combina- In short, HTML5 and WebRTC 1. Traffic and Market Data Report
tion with cloud services for real-time bring the communication landscape on the Pulse of the Networked
communication will be a very attrac- forward in two ways: they enrich Society, November 2011, http://
tive offer. The lack of guaranteed QoE communication within a community www.ericsson.com/news/120222_
for transport over the internet will be a or a vertical application such as e-health, interim_traffic_and_market_
key business driver for many enterpris- education or asocial network, and they data_report_covers_growth_in_
es and service providers. Customized enable faster deployment of mass- subscriptions_voice_traffic_and_
transport offerings that guarantee QoS market communication services. mobile_data_244159020_c

E r i c s s o n r e v i e w • 1 2012