One document matched: draft-ietf-rtcweb-security-06.xml
<?xml version="1.0" encoding="US-ASCII"?>
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
<!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml">
<!ENTITY RFC3552 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3552.xml">
<!ENTITY RFC3552 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3552.xml">
<!ENTITY RFC2818 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2818.xml">
<!ENTITY RFC3261 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3261.xml">
<!ENTITY RFC3711 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3711.xml">
<!ENTITY RFC5479 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5479.xml">
<!ENTITY RFC4347 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4347.xml">
<!ENTITY RFC4568 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4568.xml">
<!ENTITY RFC5763 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5763.xml">
<!ENTITY RFC4251 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4251.xml">
<!ENTITY RFC3760 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3760.xml">
<!ENTITY RFC6189 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6189.xml">
<!ENTITY RFC5245 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5245.xml">
<!ENTITY RFC6454 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6454.xml">
<!ENTITY RFC6455 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6455.xml">
<!ENTITY RFC6222 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6222.xml">
<!ENTITY I-D.kaufman-rtcweb-security-ui SYSTEM "http://xml.resource.org/public/rfc/bibxml3/reference.I-D.kaufman-rtcweb-security-ui.xml">
<!ENTITY I-D.muthu-behave-consent-freshness SYSTEM "http://xml.resource.org/public/rfc/bibxml3/reference.I-D.muthu-behave-consent-freshness.xml">
<!ENTITY I-D.ietf-rtcweb-security-arch SYSTEM "http://xml.resource.org/public/rfc/bibxml3/reference.I-D.ietf-rtcweb-security-arch.xml">
<!ENTITY I-D.ietf-rtcweb-overview SYSTEM "http://xml.resource.org/public/rfc/bibxml3/reference.I-D.ietf-rtcweb-overview.xml">
<!ENTITY I-D.ietf-avtcore-6222bis SYSTEM "http://xml.resource.org/public/rfc/bibxml3/reference.I-D.ietf-avtcore-6222bis.xml">
]>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
<?rfc toc="yes" ?>
<?rfc symrefs="yes" ?>
<?rfc strict="yes" ?>
<?rfc compact="yes" ?>
<?rfc sortrefs="yes" ?>
<?rfc colonspace="yes" ?>
<?rfc rfcedstyle="no" ?>
<!-- Don't change this. It breaks stuff -->
<?rfc tocdepth="4"?>
<rfc category="std" docName="draft-ietf-rtcweb-security-06"
ipr="pre5378Trust200902">
<front>
<title abbrev="WebRTC Security">Security Considerations for WebRTC</title>
<author fullname="Eric Rescorla" initials="E.K." surname="Rescorla">
<organization>RTFM, Inc.</organization>
<address>
<postal>
<street>2064 Edgewood Drive</street>
<city>Palo Alto</city>
<region>CA</region>
<code>94303</code>
<country>USA</country>
</postal>
<phone>+1 650 678 2350</phone>
<email>ekr@rtfm.com</email>
</address>
</author>
<date day="22" month="January" year="2014" />
<area>RAI</area>
<workgroup>RTC-Web</workgroup>
<abstract>
<t>
The Real-Time Communications on the Web (RTCWEB) working group is tasked with
standardizing protocols for real-time communications between Web browsers, generally
called "WebRTC". The
major use cases for WebRTC technology are real-time audio and/or video calls,
Web conferencing, and direct data transfer. Unlike most conventional real-time systems
(e.g., SIP-based soft phones) WebRTC communications are directly controlled
by a Web server, which poses new security challenges.
For instance, a Web browser might expose a JavaScript
API which allows a server to place a video call. Unrestricted access to such
an API would allow any site which a user visited to "bug" a user's computer,
capturing any activity which passed in front of their camera. This document
defines the WebRTC threat model and analyzes the security threats of
WebRTC in that model.
</t>
</abstract>
</front>
<middle>
<section title="Introduction" anchor="sec.introduction">
<t>
The Real-Time Communications on the Web (RTCWEB) working group is tasked with
standardizing protocols for real-time communications between Web browsers, generally
called "WebRTC" <xref target="I-D.ietf-rtcweb-overview"/>. The
major use cases for WebTC technology are real-time audio and/or video calls,
Web conferencing, and direct data transfer. Unlike most conventional real-time systems,
(e.g., SIP-based<xref target="RFC3261"></xref> soft phones) WebRTC communications are directly controlled
by some Web server. A simple case is shown below.
</t>
<figure title="A simple WebRTC system" anchor="fig.simple">
<artwork><![CDATA[
+----------------+
| |
| Web Server |
| |
+----------------+
^ ^
/ \
HTTP / \ HTTP
or / \ or
WebSockets / \ WebSockets
v v
JS API JS API
+-----------+ +-----------+
| | Media | |
| Browser |<---------->| Browser |
| | | |
+-----------+ +-----------+
]]></artwork>
</figure>
<t>
In the system shown in <xref target="fig.simple"/>, Alice and Bob both have
WebRTC enabled browsers and they visit some Web server which operates a
calling service. Each of their browsers exposes standardized JavaScript
calling APIs (implementated as browser built-ins)
which are used by the Web server to set up a call between Alice and Bob.
The Web server also serves as the signaling channel to transport
control messages between the browsers.
While this system is topologically similar to a conventional SIP-based
system (with the Web server acting as the signaling service and browsers
acting as softphones), control has moved to the central Web server;
the browser simply provides API points that are used by the calling service.
As with any Web application, the Web server can move logic between
the server and JavaScript in the browser, but regardless of where the
code is executing, it is ultimately under control of the server.
</t>
<t>
It should be immediately apparent that this type of system poses new
security challenges beyond those of a conventional VoIP system. In particular,
it needs to contend with malicious calling services.
For example, if the calling service
can cause the browser to make a call at any time to any callee of its
choice, then this facility can be used to bug a user's computer without
their knowledge, simply by placing a call to some recording service.
More subtly, if the exposed APIs allow the server to instruct the
browser to send arbitrary content, then they can be used to bypass
firewalls or mount denial of service attacks. Any successful system
will need to be resistant to this and other attacks.
</t>
<t>
A companion document <xref target="I-D.ietf-rtcweb-security-arch"/> describes a security
architecture intended to address the issues raised in this document.
</t>
</section>
<section anchor="sec-term" title="Terminology">
<t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in <xref
target="RFC2119">RFC 2119</xref>.</t>
</section>
<section title="The Browser Threat Model" anchor="sec.web-security">
<t>
The security requirements for WebRTC follow directly from the
requirement that the browser's job is to protect the user.
Huang et al. <xref target="huang-w2sp"/> summarize the core browser security guarantee as:
</t>
<t>
<list style="hanging">
<t>
Users can safely visit arbitrary web sites and execute scripts provided by those sites.
</t>
</list>
</t>
<t></t>
<t>
It is important to realize that this includes sites hosting arbitrary malicious
scripts. The motivation for this requirement is simple: it is trivial for attackers
to divert users to sites of their choice. For instance, an attacker can purchase
display advertisements which direct the user (either automatically or via user
clicking) to their site, at which point the browser will execute the attacker's
scripts. Thus, it is important that it be safe to view arbitrarily malicious pages.
Of course, browsers inevitably have bugs which cause them to fall short of this
goal, but any new WebRTC functionality must be designed with the intent to
meet this standard. The remainder of this section provides more background
on the existing Web security model.
</t>
<t>
In this model, then, the browser acts as a TRUSTED COMPUTING BASE (TCB) both
from the user's perspective and to some extent from the server's. While HTML
and JavaScript (JS) provided by the server can cause the browser to execute a variety of
actions, those scripts operate in a sandbox that isolates them both from
the user's computer and from each other, as detailed below.
</t>
<t>
Conventionally, we refer to either WEB ATTACKERS, who are able to induce
you to visit their sites but do not control the network, and NETWORK
ATTACKERS, who are able to control your network. Network attackers correspond
to the <xref target="RFC3552"/> "Internet Threat Model". Note that
for HTTP traffic, a network attacker is also a Web attacker,
since it can inject traffic as if it were any non-HTTPS Web
site. Thus, when analyzing HTTP connections, we must assume
that traffic is going to the attacker.
</t>
<section title="Access to Local Resources" anchor="sec.resources">
<t>
While the browser has access to local resources such as keying material,
files, the camera and the microphone, it strictly limits or forbids web
servers from accessing those same resources. For instance, while it is possible
to produce an HTML form which will allow file upload, a script cannot do
so without user consent and in fact cannot even suggest a specific file
(e.g., /etc/passwd); the user must explicitly select the file and consent
to its upload. [Note: in many cases browsers are explicitly designed to
avoid dialogs with the semantics of "click here to screw yourself", as
extensive research shows that users are prone to consent under such
circumstances.]
</t>
<t>
Similarly, while Flash programs (SWFs) <xref target="SWF"/> can access the camera and microphone, they
explicitly require that the user consent to that access. In addition,
some resources simply cannot be accessed from the browser at all. For
instance, there is no real way to run specific executables directly from a
script (though the user can of course be induced to download executable
files and run them).
</t>
</section>
<section title="Same Origin Policy" anchor="sec.same-origin">
<t>
Many other resources are accessible but isolated. For instance,
while scripts are allowed to make HTTP requests via the XMLHttpRequest() API
those requests are not allowed to be made to any server, but rather solely
to the same ORIGIN from whence the script came xref target="RFC6454"/>
(although CORS <xref target="CORS"/> and WebSockets
<xref target="RFC6455"/> provide a escape hatch from this restriction,
as described below.)
This SAME ORIGIN POLICY (SOP) prevents server A from mounting attacks
on server B via the user's browser, which protects both the user
(e.g., from misuse of his credentials) and the server B (e.g., from
DoS attack).
</t>
<t>
More generally, SOP forces scripts from each site to run in their own, isolated,
sandboxes. While there are techniques to allow them to interact, those interactions
generally must be mutually consensual (by each site) and are limited to certain
channels. For instance, multiple pages/browser panes from the same origin
can read each other's JS variables, but pages from the different origins--or
even iframes from different origins on the same page--cannot.
</t>
<!-- TODO: Picture -->
</section>
<section title="Bypassing SOP: CORS, WebSockets, and consent to communicate" anchor="sec.cors-etc">
<t>
While SOP serves an important security function, it also makes it inconvenient to
write certain classes of applications. In particular, mash-ups, in which a script
from origin A uses resources from origin B, can only be achieved via a certain amount of hackery.
The W3C Cross-Origin Resource Sharing (CORS) spec <xref target="CORS"/> is a response to this
demand. In CORS, when a script from origin A executes what would otherwise be a forbidden
cross-origin request, the browser instead contacts the target server to determine
whether it is willing to allow cross-origin requests from A. If it is so willing,
the browser then allows the request. This consent verification process is designed
to safely allow cross-origin requests.
</t>
<t>
While CORS is designed to allow cross-origin HTTP requests, WebSockets <xref target="RFC6455"/> allows
cross-origin establishment of transparent channels. Once a WebSockets connection
has been established from a script to a site, the script can exchange any traffic it
likes without being required to frame it as a series of HTTP request/response
transactions. As with CORS, a WebSockets transaction starts with a consent verification
stage to avoid allowing scripts to simply send arbitrary data to another origin.
</t>
<t>
While consent verification is conceptually simple--just do a handshake before you
start exchanging the real data--experience has shown that designing a
correct consent verification system is difficult. In particular, Huang et al. <xref target="huang-w2sp"/>
have shown vulnerabilities in the existing Java and Flash consent verification
techniques and in a simplified version of the WebSockets handshake. In particular,
it is important to be wary of CROSS-PROTOCOL attacks in which the attacking script
generates traffic which is acceptable to some non-Web protocol state machine.
In order to resist this form of attack, WebSockets incorporates a masking technique
intended to randomize the bits on the wire, thus making it more difficult to generate
traffic which resembles a given protocol.
</t>
</section>
</section>
<section title="Security for WebRTC Applications" anchor="sec.rtc-web">
<section title="Access to Local Devices" anchor="sec.rtc-dev-access">
<t>
As discussed in <xref target="sec.introduction"/>, allowing arbitrary
sites to initiate calls violates the core Web security guarantee;
without some access restrictions on local devices, any malicious site
could simply bug a user. At minimum, then, it MUST NOT be possible for
arbitrary sites to initiate calls to arbitrary locations without user
consent. This immediately raises the question, however, of what should
be the scope of user consent.
</t>
<t>
In order for the user to
make an intelligent decision about whether to allow a call
(and hence his camera and microphone input to be routed somewhere),
he must understand either who is requesting access, where the media
is going, or both. As detailed below, there are two basic conceptual
models:
</t>
<t>
<list>
<t>You are sending your media to entity A because you want to
talk to Entity A (e.g., your mother).</t>
<t>Entity A (e.g., a calling service) asks to access the user's devices with the assurance
that it will transfer the media to entity B (e.g., your mother)</t>
</list>
</t>
<t>
In either case, identity is at the heart of any consent decision.
Moreover, identity is all that the browser can meaningfully enforce;
if you are calling A, A can simply forward the media to C. Similarly,
if you authorize A to place a call to B, A can call C instead.
In either case, all the browser is able to do is verify and check
authorization for whoever is controlling where the media goes.
The target of the media can of course advertise a security/privacy
policy, but this is not something that the browser can
enforce. Even so, there are a variety of different consent scenarios
that motivate different technical consent mechanisms.
We discuss these mechanisms in the sections below.
</t>
<t>
It's important to understand that consent to access local devices
is largely orthogonal to consent to transmit various kinds of
data over the network (see <xref target="sec.rtc-comm-consent"/>.
Consent for device access is largely a matter of protecting
the user's privacy from malicious sites. By contrast,
consent to send network traffic is about preventing the
user's browser from being used to attack its local network.
Thus, we need to ensure communications consent even if the
site is not able to access the camera and microphone at
all (hence WebSockets's consent mechanism) and similarly
we need to be concerned with the site accessing the
user's camera and microphone even if the data is to be
sent back to the site via conventional HTTP-based network
mechanisms such as HTTP POST.
</t>
<section title="Threats from Screen Sharing">
<t>
In addition to camera and microphone access, there has been
demand for screen and/or application sharing functionality.
Unfortunately, the security implications of this
functionality are much harder for users to intuitively
analyze than for camera and microphone access.
(See http://lists.w3.org/Archives/Public/public-webrtc/2013Mar/0024.html
for a full analysis.)
</t>
<t>
The most obvious threats are simply those of "oversharing".
I.e., the user may believe they are sharing a window when
in fact they are sharing an application, or may forget they
are sharing their whole screen, icons, notifications, and all.
This is already an issue with existing screen sharing technologies
and is made somewhat worse if a partially trusted site is responsible for asking
for the resource to be shared rather than having the user propose it.
</t>
<t>
A less obvious threat involves the impact of screen sharing on the
Web security model. A key part of the Same Origin Policy is that
HTML or JS from site A can reference content from site B and cause
the browser to load it, but (unless explicitly permitted) cannot
see the result. However, if a web application from a site is
screen sharing the browser, then this violates that invariant,
with serious security consequences. For example, an attacker site
might request screen sharing and then briefly open up a new
Window to the user's bank or webmail account, using screen sharing
to read the resulting displayed content. A more sophisticated
attack would be open up a source view window to a site and use the
screen sharing result to view anti cross-site request forgery tokens.
</t>
<t>
These threats suggest that screen/application sharing might need
a higher level of user consent than access to the camera or
microphone.
</t>
</section>
<section title="Calling Scenarios and User Expectations">
<t>
While a large number of possible calling scenarios are possible, the
scenarios discussed in this section illustrate many of
the difficulties of identifying the relevant scope of consent.
</t>
<section title="Dedicated Calling Services">
<t>
The first scenario we consider is a dedicated calling service. In this
case, the user has a relationship with a calling site
and repeatedly makes calls on it. It is likely
that rather than having to give permission for each call
that the user will want to give the calling service long-term
access to the camera and microphone. This is a natural fit
for a long-term consent mechanism (e.g., installing an
app store "application" to indicate permission for the
calling service.)
A variant of the dedicated calling service is a gaming site
(e.g., a poker site) which hosts a dedicated calling service
to allow players to call each other.
</t>
<t>
With any kind of service where the user may use the same
service to talk to many different people, there is a question
about whether the user can know who they are talking to.
If I grant permission to calling service A to make calls
on my behalf, then I am implicitly granting it permission
to bug my computer whenever it wants. This suggests another
consent model in which a site is authorized to make calls
but only to certain target entities (identified via
media-plane cryptographic mechanisms as described in
<xref target="sec.during-attack"/> and especially
<xref target="sec.third-party-id"/>.) Note that the
question of consent here is related to but
distinct from the question of peer identity: I
might be willing to allow a calling site to in general
initiate calls on my behalf but still have some calls
via that site where I can be sure that the site is not
listening in.
</t>
</section>
<section title="Calling the Site You're On">
<t>
Another simple scenario is calling the site you're actually visiting.
The paradigmatic case here is the "click here to talk to a
representative" windows that appear on many shopping sites.
In this case, the user's expectation is that they are
calling the site they're actually visiting. However, it is
unlikely that they want to provide a general consent to such
a site; just because I want some information on a car
doesn't mean that I want the car manufacturer to be able
to activate my microphone whenever they please. Thus,
this suggests the need for a second consent mechanism
where I only grant consent for the duration of a given
call. As described in <xref target="sec.resources"/>,
great care must be taken in the design of this interface
to avoid the users just clicking through. Note also
that the user interface chrome must clearly display elements
showing that the call is continuing in order to avoid attacks
where the calling site just leaves it up indefinitely but
shows a Web UI that implies otherwise.
</t>
</section>
</section>
<section title="Origin-Based Security">
<t>
Now that we have seen another use case, we can start to reason about
the security requirements.
</t>
<t>
As discussed in <xref target="sec.same-origin"/>, the basic unit of
Web sandboxing is the origin, and so it is natural to scope consent
to origin. Specifically, a script from origin A MUST only be allowed
to initiate communications (and hence to access camera and microphone)
if the user has specifically authorized access for that origin.
It is of course technically possible to have coarser-scoped permissions,
but because the Web model is scoped to origin, this creates a difficult
mismatch.
</t>
<t>
Arguably, origin is not fine-grained enough. Consider the situation where
Alice visits a site and authorizes it to make a single call. If consent is
expressed solely in terms of origin, then at any future visit to that
site (including one induced via mash-up or ad network), the site can
bug Alice's computer, use the computer to place bogus calls, etc.
While in principle Alice could grant and then
revoke the privilege, in practice privileges accumulate; if we are concerned
about this attack, something else is needed. There are a number of potential countermeasures to
this sort of issue.
</t>
<t><list style="hanging">
<t hangText="Individual Consent"></t><t>Ask the user for permission for each call.</t>
<t></t>
<t hangText="Callee-oriented Consent"></t><t>Only allow calls to a given user.</t>
<t></t>
<t hangText="Cryptographic Consent"></t><t>Only allow calls to a given set of peer keying material or
to a cryptographically established identity.</t>
</list>
</t>
<t>
Unfortunately, none of these approaches is satisfactory for all cases.
As discussed above, individual consent puts the user's approval
in the UI flow for every call. Not only does this quickly become annoying
but it can train the user to simply click "OK", at which point the consent becomes
useless. Thus, while it may be necessary to have individual consent in some
case, this is not a suitable solution for (for instance) the calling
service case. Where necessary, in-flow user interfaces must be carefully
designed to avoid the risk of the user blindly clicking through.
</t>
<t>
The other two options are designed to restrict calls to a given target.
Callee-oriented consent provided by the calling site
not work well because a malicious site can claim that the
user is calling any user of his choice. One fix for this is to tie calls to a
cryptographically established identity. While not suitable for all cases,
this approach may be useful for some. If we consider the case
of advertising, it's not particularly convenient
to require the advertiser to instantiate an iframe on the hosting site just
to get permission; a more convenient approach is to cryptographically tie
the advertiser's certificate to the communication directly. We're still
tying permissions to origin here, but to the media origin (and-or destination)
rather than to the Web origin. <xref target="I-D.ietf-rtcweb-security-arch"/>
describes mechanisms
which facilitate this sort of consent.
</t>
<t>
Another case where media-level cryptographic identity makes sense is when a user
really does not trust the calling site. For instance, I might be worried that
the calling service will attempt to bug my computer, but I also want to be
able to conveniently call my friends. If consent is tied to particular
communications endpoints, then my risk is limited. Naturally, it
is somewhat challenging to design UI primitives which express this sort
of policy. The problem becomes even more challenging in multi-user
calling cases.
</t>
</section>
<section title="Security Properties of the Calling Page">
<t>
Origin-based security is intended to secure against web attackers. However, we must
also consider the case of network attackers. Consider the case where I have
granted permission to a calling service by an origin that has the HTTP scheme,
e.g., http://calling-service.example.com. If I ever use my computer on
an unsecured network (e.g., a hotspot or if my own home wireless network
is insecure), and browse any HTTP site, then an attacker can bug my computer. The attack proceeds
like this:
</t>
<t>
<list style="numbers">
<t>I connect to http://anything.example.org/. Note that this site is unaffiliated
with the calling service.</t>
<t>The attacker modifies my HTTP connection to inject an IFRAME (or a redirect)
to http://calling-service.example.com</t>
<t>The attacker forges the response apparently http://calling-service.example.com/ to
inject JS to initiate a call to himself.</t>
</list>
</t>
<t>
Note that this attack does not depend on the media being insecure. Because the
call is to the attacker, it is also encrypted to him. Moreover, it need not
be executed immediately; the attacker can "infect" the origin semi-permanently
(e.g., with a web worker or a popped-up window that is hidden under the main window.)
and thus be able to bug me long
after I have left the infected network. This risk is created by allowing
calls at all from a page fetched over HTTP.
</t>
<t>
Even if calls are only possible from HTTPS sites, if the
site embeds active content (e.g., JavaScript) that is fetched over
HTTP or from an untrusted site, because that JavaScript is executed
in the security context of the page <xref target="finer-grained"/>.
Thus, it is also dangerous to allow WebRTC functionality from
HTTPS origins that embed mixed content.
Note: this issue is not restricted
to PAGES which contain mixed content. If a page from a given origin ever loads mixed content
then it is possible for a network attacker to infect the browser's notion of that
origin semi-permanently.
</t>
</section>
</section>
<section title="Communications Consent Verification" anchor="sec.rtc-comm-consent">
<t>
As discussed in <xref target="sec.cors-etc"/>, allowing web applications unrestricted network access
via the browser introduces the risk of using the browser as an attack platform against
machines which would not otherwise be accessible to the malicious site, for
instance because they are topologically restricted (e.g., behind a firewall or NAT).
In order to prevent this form of attack as well as cross-protocol attacks it is
important to require that the target of traffic explicitly consent to receiving
the traffic in question. Until that consent has been verified for a given endpoint,
traffic other than the consent handshake MUST NOT be sent to that endpoint.
</t>
<section title="ICE" anchor="sec.ice">
<t>
Verifying receiver consent requires some sort of explicit handshake, but conveniently
we already need one in order to do NAT hole-punching. ICE <xref target="RFC5245"/> includes a handshake
designed to verify that the receiving element wishes to receive traffic from the
sender. It
is important to remember here that the site initiating ICE is
presumed malicious; in order for the handshake to be secure the
receiving element MUST demonstrate receipt/knowledge of some value
not available to the site (thus preventing the site from forging
responses). In order to achieve this objective with ICE, the STUN
transaction IDs must be generated by the browser and MUST NOT be made
available to the initiating script, even via a diagnostic interface.
Verifying receiver consent also requires verifying the receiver wants
to receive traffic from a particular sender, and at this time; for
example a malicious site may simply attempt ICE to known servers
that are using ICE for other sessions. ICE provides this verification
as well, by using the STUN credentials as a form of per-session shared
secret. Those credentials are known to the Web application, but would
need to also be known and used by the STUN-receiving element to be useful.
</t>
<t>
There also needs to be some mechanism for the browser to verify that
the target of the traffic continues to wish to receive it. Because ICE keepalives are
indications, they will not work here, so some other mechanism is
needed as described in <xref target="I-D.muthu-behave-consent-freshness"/>.
</t>
</section>
<section title="Masking" anchor="sec.masking">
<t>
Once consent is verified, there still is some concern about misinterpretation
attacks as described by Huang et al.<xref target="huang-w2sp"/>.
Once consent is verified, there still is some concern about
misinterpretation attacks as described by Huang et al.[huang-w2sp].
Where TCP is used the risk is substantial due to the potential
presence of transparent proxies and therefore if TCP is to be used,
then WebSockets style masking MUST be employed.
</t>
<t>
Since DTLS (with the anti-chosen plaintext mechanisms required by
TLS 1.1) does not allow the attacker to generate predictable
ciphertext, there is no need for masking of protocols running over
DTLS (e.g. SCTP over DTLS, UDP over DTLS, etc.).
</t>
</section>
<section title="Backward Compatibility">
<t>
A requirement to use ICE limits compatibility with legacy non-ICE clients.
It seems unsafe to completely remove the requirement for some check.
All proposed checks have the common feature that the browser
sends some message to the candidate traffic recipient
and refuses to send other traffic until that message has been
replied to. The message/reply pair must be generated in such
a way that an attacker who controls the Web application
cannot forge them, generally by having the message contain some
secret value that must be incorporated (e.g., echoed, hashed into,
etc.). Non-ICE candidates for this role (in cases where the
legacy endpoint has a public address) include:
</t>
<t>
<list style="symbols">
<t>STUN checks without using ICE (i.e., the non-RTC-web endpoint sets up a STUN responder.)</t>
<t>Use or RTCP as an implicit reachability check.</t>
</list>
</t>
<t>
In the RTCP approach, the WebRTC endpoint is allowed to send
a limited number of RTP packets prior to receiving consent. This
allows a short window of attack. In addition, some legacy endpoints
do not support RTCP, so this is a much more expensive solution for
such endpoints, for which it would likely be easier to implement ICE.
For these two reasons, an RTCP-based approach does not seem to
address the security issue satisfactorily.
</t>
<t>
In the STUN approach, the WebRTC endpoint is able to verify that
the recipient is running some kind of STUN endpoint but unless
the STUN responder is integrated with the ICE username/password
establishment system, the WebRTC endpoint cannot verify that
the recipient consents to this particular call. This may be an
issue if existing STUN servers are operated at addresses that
are not able to handle bandwidth-based attacks. Thus, this
approach does not seem satisfactory either.
</t>
<t>
If the systems are tightly integrated (i.e., the STUN endpoint responds with
responses authenticated with ICE credentials) then this issue
does not exist. However, such a design is very close to an ICE-Lite
implementation (indeed, arguably is one).
An intermediate approach would be to have a STUN extension that indicated
that one was responding to WebRTC checks but not computing
integrity checks based on the ICE credentials. This would allow the
use of standalone STUN servers without the risk of confusing them
with legacy STUN servers. If a non-ICE legacy solution is needed,
then this is probably the best choice.
</t>
<t>
Once initial consent is verified, we also need to verify continuing
consent, in order to avoid attacks where two people briefly share
an IP (e.g., behind a NAT in an Internet cafe) and the attacker
arranges for a large, unstoppable, traffic flow to the
network and then leaves. The appropriate technologies here are
fairly similar to those for initial consent, though are perhaps
weaker since the threats is less severe.
</t>
</section>
<section title="IP Location Privacy" anchor="sec.ip.location">
<t>
Note that as soon as the callee sends their ICE candidates, the
caller learns the callee's IP addresses. The callee's server reflexive
address reveals a lot of information about the callee's location.
In order to avoid tracking, implementations may wish to suppress
the start of ICE negotiation until the callee has answered. In
addition, either side may wish to hide their location entirely
by forcing all traffic through a TURN server.
</t>
<t>
In ordinary operation, the site learns the browser's IP address,
though it may be hidden via mechanisms like Tor [http://www.torproject.org] or a VPN.
However, because sites can cause the browser to provide
IP addresses, this provides a mechanism for sites to learn
about the user's network environment even if the user is behind
a VPN that masks their IP address. Implementations wish
to provide settings which suppress all non-VPN candidates if
the user is on certain kinds of VPN, especially privacy-oriented
systems such as Tor.
</t>
</section>
</section>
<section title="Communications Security" anchor="sec.rtc-comsec">
<t>
Finally, we consider a problem familiar from the SIP world: communications security.
For obvious reasons, it MUST be possible for the communicating parties to establish
a channel which is secure against both message recovery and message modification.
(See <xref target="RFC5479"/> for more details.)
This service must be provided for both data and voice/video.
Ideally the same security mechanisms would be used for both types of content.
Technology for providing this
service (for instance, SRTP <xref target="RFC3711"/>, DTLS <xref target="RFC4347"/> and
DTLS-SRTP <xref target="RFC5763"/>) is well understood. However, we must
examine this technology to the WebRTC context, where the threat
model is somewhat different.
</t>
<t>
In general, it is important to understand that unlike a conventional SIP proxy,
the calling service (i.e., the Web server) controls not only the channel
between the communicating endpoints but also the application running on
the user's browser.
While in principle it is possible for the browser to cut the calling service
out of the loop and directly present trusted information (and perhaps get
consent), practice in modern browsers is to avoid this whenever possible.
"In-flow" modal dialogs which require the user to consent to specific
actions are particularly disfavored as human factors research indicates
that unless they are made extremely invasive, users simply agree to
them without actually consciously giving consent. <xref target="abarth-rtcweb"/>.
Thus, nearly all the UI will necessarily be rendered by the
browser but under control of the calling service. This likely includes the
peer's identity information, which, after all, is only meaningful in
the context of some calling service.
</t>
<t>
This limitation does not mean that preventing attack by the calling service
is completely hopeless. However, we need to distinguish between two
classes of attack:
</t>
<t><list style="hanging">
<t hangText="Retrospective compromise of calling service."></t><t>The calling service is
is non-malicious during a call but subsequently is compromised and wishes to
attack an older call (often called a "passive attack")</t>
<t></t>
<t hangText="During-call attack by calling service."></t><t>The calling service is compromised
during the call it wishes to attack (often called an "active attack").</t>
</list>
</t>
<t>
Providing security against the former type of attack is practical using the
techniques discussed in <xref target="sec.retrospective-compromise"/>.
However, it is extremely difficult to prevent a
trusted but malicious calling service from actively attacking a user's calls,
either by mounting a MITM attack or by diverting them entirely.
(Note that this attack applies equally to a network attacker if communications
to the calling service are not secured.) We discuss some potential approaches
and why they are likely to be impractical in <xref target="sec.during-attack"/>.
</t>
<section title="Protecting Against Retrospective Compromise" anchor="sec.retrospective-compromise">
<t>
In a retrospective attack, the calling service was uncompromised during
the call, but that an attacker subsequently wants to recover the content of the
call. We assume that the attacker has access to the protected media stream
as well as having full control of the calling service.
</t>
<t>
If the calling service has access to the traffic keying material
(as in SDES <xref target="RFC4568"/>), then retrospective attack
is trivial.
This form of attack is particularly serious in the Web context because
it is standard practice in Web services to run extensive logging and monitoring. Thus, it is highly
likely that if the traffic key is part of any HTTP request it will be logged somewhere and thus
subject to subsequent compromise. It is this consideration that makes an automatic, public key-based
key exchange mechanism imperative for WebRTC (this is a good idea for any communications
security system) and this mechanism SHOULD provide perfect forward secrecy (PFS).
The signaling channel/calling service can be used to authenticate this mechanism.
</t>
<t>
In addition, if end-to-end keying is in used,
the system MUST NOT provide any APIs to extract either long-term
keying material or to directly access any stored traffic keys.
Otherwise, an attacker who subsequently compromised the calling service
might be able to use those APIs to recover the traffic keys and thus
compromise the traffic.
</t>
</section>
<section title="Protecting Against During-Call Attack" anchor="sec.during-attack">
<t>
Protecting against attacks during a call is a more difficult proposition. Even
if the calling service cannot directly access keying material (as recommended
in the previous section), it can simply mount a man-in-the-middle attack
on the connection, telling Alice that she is calling Bob and Bob that
he is calling Alice, while in fact the calling service is acting as
a calling bridge and capturing all the traffic. Protecting against
this form of attack requires positive authentication of the remote
endpoint such as explicit out-of-band key verification (e.g., by a fingerprint)
or a third-party identity service as described in
<xref target="I-D.ietf-rtcweb-security-arch"/>.
</t>
<section title="Key Continuity" anchor="sec.key-continuity">
<t>
One natural approach is to use "key continuity". While a malicious
calling service can present any identity it chooses to the user,
it cannot produce a private key that maps to a given public key.
Thus, it is possible for the browser to note a given user's
public key and generate an alarm whenever that user's key
changes. SSH <xref target="RFC4251"/> uses a similar technique.
(Note that the need to avoid explicit user consent on every call
precludes the browser requiring an immediate manual check of the peer's key).
</t>
<t>
Unfortunately, this sort of key continuity mechanism is far less
useful in the WebRTC context. First, much of the virtue of
WebRTC (and any Web application) is that it is not bound to
particular piece of client software. Thus, it will be not only
possible but routine for a user to use multiple browsers
on different computers which will of course have different
keying material (SACRED <xref target="RFC3760"/> notwithstanding.)
Thus, users will frequently be alerted to key mismatches which
are in fact completely legitimate, with the result that they
are trained to simply click through them. As it is known that
users routinely will click through far more dire warnings
<xref target="cranor-wolf"/>, it seems extremely unlikely that
any key continuity mechanism will be effective rather than
simply annoying.
</t>
<t>
Moreover, it is trivial to bypass even this kind of mechanism.
Recall that unlike the case of SSH, the browser never directly
gets the peer's identity from the user. Rather, it is provided
by the calling service. Even enabling a mechanism of this type
would require an API to allow the calling service to tell the
browser "this is a call to user X". All the calling service
needs to do to avoid triggering a key continuity warning
is to tell the browser that "this is a call to user Y"
where Y is close to X.
Even if the user actually checks the other side's name
(which all available evidence indicates is unlikely),
this would require (a) the browser to trusted UI
to provide the name and (b) the user to not be fooled by
similar appearing names.
</t>
</section>
<section title="Short Authentication Strings" anchor="sec.sas">
<t>
ZRTP <xref target="RFC6189"/> uses a "short authentication string" (SAS) which is derived
from the key agreement protocol. This SAS is designed to be compared
by the users (e.g., read aloud over the the voice channel or
transmitted via an out of band channel) and if confirmed by both sides precludes MITM
attack. The intention is that the SAS is used once and then key
continuity (though a different mechanism from that discussed
above) is used thereafter.
</t>
<t>
Unfortunately, the SAS does not offer a practical solution to the
problem of a compromised calling service. "Voice conversion" systems, which modify
voice from one speaker to make it sound like another,
are an active area of research.
These systems are already good enough to fool both
automatic recognition systems <xref target="farus-conversion"/> and
humans <xref target="kain-conversion"/> in many cases, and are of course likely
to improve in future, especially in an environment where the user just wants
to get on with the phone call.
Thus, even if SAS is effective today, it is likely not to be so for much longer.
</t>
<t>
Additionally, it is unclear that users will actually use an SAS.
As discussed above, the browser UI constraints preclude requiring
the SAS exchange prior to completing the call and so it must be
voluntary; at most the browser will provide some UI indicator that the
SAS has not yet been checked. However, it it is well-known that when
faced with optional security mechanisms, many users simply
ignore them <xref target="whitten-johnny"/>.
</t>
<t>
Once uses have checked the SAS once, key continuity
is required to avoid them needing to check it on every call.
However, this is problematic for reasons indicated in
<xref target="sec.key-continuity"/>.
In principle it is of course possible to render a different
UI element to indicate that calls are using an unauthenticated
set of keying material (recall that the attacker can just present
a slightly different name so that the attack shows the
same UI as a call to a new device or to someone you haven't
called before) but as a practical matter, users simply ignore
such indicators even in the rather more dire case of mixed
content warnings.
</t>
</section>
<section title="Third Party Identity" anchor="sec.third-party-id">
<t>
The conventional approach to providing communications identity
has of course been to have some third party identity system
(e.g., PKI) to authenticate the endpoints. Such mechanisms
have proven to be too cumbersome for use by typical users
(and nearly too cumbersome for administrators).
However,
a new generation of Web-based identity providers (BrowserID, Federated Google Login,
Facebook Connect, OAuth, OpenID, WebFinger), has recently been developed
and use Web technologies to provide lightweight (from the user's
perspective) third-party authenticated transactions.
It is possible to use systems of this type to authenticate WebRTC calls,
linking them to existing user notions of identity
(e.g., Facebook adjacencies). Specifically, the third-party
identity system is used to bind the user's identity to
cryptographic keying material which is then used to
authenticate the calling endpoints.
Calls which are authenticated
in this fashion are naturally resistant even to active MITM attack
by the calling site.
</t>
<t>
Note that there is one special case in which PKI-style certificates
do provide a practical solution: calls from end-users to
large sites. For instance, if you are making a call
to Amazon.com, then Amazon can easily get a certificate
to authenticate their media traffic, just as they get
one to authenticate their Web traffic. This does not provide
additional security value in cases in which the calling site
and the media peer are one in the same, but might be useful
in cases in which third parties (e.g., ad networks or
retailers) arrange for calls but do not participate in them.
</t>
</section>
<section title="Page Access to Media" anchor="sec.page-access">
<t>
Identifying the identity of the far media endpoint is a
necessary but not sufficient condition for providing media
security. In WebRTC, media flows are rendered into
HTML5 MediaStreams which can be manipulated by the calling
site. Obviously, if the site can modify or view the media,
then the user is not getting the level of assurance they
would expect from being able to authenticate their peer.
In many cases, this is acceptable because the user values
site-based special effects over complete security from the
site. However, there are also cases where users wish to
know that the site cannot interfere. In order to facilitate
that, it will be necessary to provide features whereby
the site can verifiably give up access to the media streams.
This verification must be possible both from the local
side and the remote side. I.e., I must be able to verify
that the person I am calling has engaged a secure media
mode. In order to achieve this it will be necessary to
cryptographically bind an indication of the local media
access policy into the cryptographic authentication
procedures detailed in the previous sections.
</t>
</section>
</section>
<section title="Malicious Peers" anchor="sec.malicious">
<t>
One class of attack that we do not generally try to prevent
is malicious peers. For instance, no matter what confidentiality
measures you employ the person you are talking to might record
the call and publish it on the Internet. Similarly, we do
not attempt to prevent them from using voice or video processing
technology from hiding or changing their appearance.
While technologies (DRM, etc.) do exist to attempt to address
these issues, they are generally not compatible with open
systems and WebRTC does not address them.
</t>
<t>
Similarly, we make no attempt to prevent prank calling or
other unwanted calls. In general, this is in the scope of the
calling site, though because WebRTC does offer some forms of
strong authentication, that may be useful as part of a defense
against such attacks.
</t>
</section>
</section>
<section title="Privacy Considerations" anchor="sec.privacy">
<section title="Correlation of Anonymous Calls">
<t>
While persistent endpoint identifiers can be a useful security
feature (see <xref target="sec.key-continuity"/> they can
also represent a privacy threat in settings where the user
wishes to be anonymous. WebRTC provides a number of possible
persistent identifiers such as DTLS certificates
(if they are reused between connections) and RTCP CNAMES
(if generated according to <xref target="RFC6222"/> rather
than the privacy preserving mode of <xref target="I-D.ietf-avtcore-6222bis"/>).
In order to prevent this type of correlation, browsers need to
provide mechanisms to reset these identifiers (e.g., with the
same lifetime as cookies). Moreover, the API should provide
mechanisms to allow sites intended for anonymous calling
to force the minting of fresh identifiers.
</t>
</section>
<section title="Browser Fingerprinting">
<t>
Any new set of API features adds a risk of browser fingerprinting,
and WebRTC is no exception. Specifically, sites can use the
presence or absence of specific devices as a browser fingerprint.
In general, the API needs to be balanced between functionality
and the incremental fingerprint risk.
</t>
</section>
</section>
</section>
<section title="Security Considerations" anchor="sec.sec_cons">
<t>This entire document is about security.</t>
</section>
<section title="Acknowledgements">
<t>
Bernard Aboba, Harald Alvestrand, Dan Druta,
Cullen Jennings, Alan Johnston, Hadriel Kaplan (S 4.2.1), Matthew Kaufman,
Martin Thomson, Magnus Westerland.
</t>
<t></t>
</section>
<section title="Changes Since -04">
<t>
<list style="symbols">
<t>Replaced RTCWEB and RTC-Web with WebRTC, except when referring to the IETF WG</t>
<t>Removed discussion of the IFRAMEd advertisement case, since we decided not to
treat it specially.</t>
<t>Added a privacy section considerations section.</t>
<t>Significant edits to the SAS section to reflect Alan Johnston's comments.</t>
<t>Added some discussion if IP location privacy and Tor.</t>
<t>Updated the "communications consent" section to reflrect draft-muthu.</t>
<t>Added a section about "malicious peers".</t>
<t>Added a section describing screen sharing threats.</t>
<t>Assorted editorial changes.</t>
</list>
</t>
</section>
</middle>
<back>
<references title="Normative References">
&RFC2119;
&I-D.ietf-rtcweb-overview;
</references>
<references title="Informative References">
&RFC3261;
&RFC3552;
&RFC3711;
&RFC2818;
&RFC5479;
&RFC5763;
&RFC4347;
&RFC4568;
&RFC4251;
&RFC3760;
&RFC6189;
&RFC5245;
&RFC6222;
&RFC6454;
&RFC6455;
&I-D.kaufman-rtcweb-security-ui;
&I-D.ietf-rtcweb-security-arch;
&I-D.ietf-avtcore-6222bis;
&I-D.muthu-behave-consent-freshness;
<reference anchor="abarth-rtcweb">
<front>
<title>Prompting the user is security failure</title>
<author initials="A." surname="Barth">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="RTC-Web Workshop"/>
<format target="http://rtc-web.alvestrand.com/home/papers/barth-security-prompt.pdf?attredirects=0" type="PDF"/>
</reference>
<reference anchor="whitten-johnny">
<front>
<title>Why Johnny Can't Encrypt: A Usability Evaluation of PGP 5.0</title>
<author initials="A." surname="Whitten">
<organization></organization>
</author>
<author initials="J.D." surname="Tygar">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="Proceedings of the 8th USENIX Security Symposium, 1999"/>
</reference>
<reference anchor="cranor-wolf">
<front>
<title>Crying Wolf: An Empirical Study of SSL Warning Effectiveness</title>
<author initials="J." surname="Sunshine">
<organization></organization>
</author>
<author initials="S." surname="Egelman">
<organization></organization>
</author>
<author initials="H." surname="Almuhimedi">
<organization></organization>
</author>
<author initials="N." surname="Atri">
<organization></organization>
</author>
<author initials="L." surname="cranor">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="Proceedings of the 18th USENIX Security Symposium, 2009"/>
</reference>
<reference anchor="kain-conversion">
<front>
<title>Design and Evaluation of a Voice Conversion Algorithm based on Spectral Envelope Mapping and Residual Prediction</title>
<author initials="A." surname="Kain">
<organization></organization>
</author>
<author initials="M." surname="Macon">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="Proceedings of ICASSP, May 2001"/>
</reference>
<reference anchor="farus-conversion">
<front>
<title>Speaker Recognition Robustness to Voice Conversion</title>
<author initials="M." surname="Farrus">
<organization></organization>
</author>
<author initials="D." surname="Erro">
<organization></organization>
</author>
<author initials="J." surname="Hernando">
<organization></organization>
</author>
</front>
</reference>
<reference anchor="huang-w2sp">
<front>
<title>Talking to Yourself for Fun and Profit</title>
<author initials="L-S." surname="Huang">
<organization></organization>
</author>
<author initials="E.Y." surname="Chen">
<organization></organization>
</author>
<author initials="A." surname="Barth">
<organization></organization>
</author>
<author initials="E." surname="Rescorla">
<organization></organization>
</author>
<author initials="C." surname="Jackson">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="W2SP, 2011"/>
</reference>
<reference anchor="finer-grained">
<front>
<title>Beware of Finer-Grained Origins</title>
<author initials="A." surname="Barth">
<organization></organization>
</author>
<author initials="C." surname="Jackson">
<organization></organization>
</author>
</front>
<seriesInfo name="" value="W2SP, 2008"/>
</reference>
<reference anchor="CORS">
<front>
<title>Cross-Origin Resource Sharing</title>
<author initials="A." surname="van Kesteren">
<organization></organization>
</author>
</front>
<format target="http://www.w3.org/TR/cors/" type="TXT"/>
</reference>
<reference anchor="SWF">
<front>
<title>SWF File Format Specification Version 19</title>
<author surname="Adobe">
<organization></organization>
</author>
</front>
<format target="http://www.adobe.com/content/dam/Adobe/en/devnet/swf/pdf/swf_file_format_spec_v10.pdf" type="PDF"/>
</reference>
</references>
</back>
<!--
On Thu, Sep 22, 2011 at 5:45 AM, Magnus Westerlund <magnus.westerlund@ericsson.com> wrote:
> Hi EKR,
>
> (As an individual)
>
> Thanks for posting the draft.
>
> I am missing a few security issues that I think should be considered.
>
> 1. The attempt to overload the links in an domain by concentrating
> traffic on the domain by choosing peer-pairs. Not that I think there is
> any real protection against this other than limit the flows to their
> "fair" share.
>
> 2. Configuring RTCP or other automatically sent traffic to high
> bit-rates. Especially under conditions where continued consent can't be
> determined.
>
> Cheers
>
> Magnus Westerlund
>
>
>
-->
<!-- drill down -->
</rfc>
| PAFTECH AB 2003-2026 | 2026-04-24 01:06:18 |