Passwords in the Clear

[Editor's Draft] TAG Finding passwordsInTheClear-52, June 02 2008

Latest version:
This version:
Previous version(s):
Previous version(s):
David Orchard, Invited Expert mailto:orchard@pacificspirit.com (and before at BEA Systems)
Previous Editor:
Ed Rice, Hewlett Packard mailto:Ed.Rice@hp.com (until Dec 2006)


The purpose of this finding is to provide guidance for securely transmitting passwords on the World Wide Web. The principle advice provided is that passwords MUST NOT be transmitted in the clear.

Status of this Document

This document has been produced by the W3C Technical Architecture Group (TAG). This finding addresses TAG issue passwordsInTheClear-52.

This finding is an editorial draft, not yet accepted by the TAG. This finding therefore does not have any official standing and does not represent the consensus of the Technical Architecture Group at the W3C.

Additional TAG findings, both accepted and in draft state, may also be available. The TAG expects to incorporate this finding according to the process of the W3C Recommendation Track.

Please send comments on this finding to the publicly archived TAG mailing list www-tag@w3.org (archive).

Table of Contents

1 Introduction
2 Passwords in the clear
    1) Secure transfers
       1) Digest Authentication
       2) SSL/TLS
    2) SOAP Based transmissions.
3 Passwords displayed in Browser
A References

1 Introduction

Security on the World Wide Web is an important issue which needs to be addressed, or mistrust of the Web will limit its growth potential. This finding describes the use of passwords on the World Wide Web and the need to keep them secure during display, temporary storage in cookies, and in transmission over the Web. Note that there are technologies other than passwords for enabling the transmission of secure informaton.

The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [ IETF RFC 2119]

The principle advice provided in this finding is that passwords MUST NOT be transmitted in the clear.

2 Passwords in the clear

This section addresses the issue of transmitting passwords in clear text over the World Wide Web. When a password is transmitted in clear text, it is vulnerable in many ways:

  1. The password is available on the wire. As the password is transmitted over the wire, tools such as packet sniffers or network analyzers can easily monitor the traffic and intercept passwords as they're sent between computers.
  2. The password is available in browsing history. Most web browsers provide 'back' navigation to previous pages, with content locally cached for performance as well as ease of use for the user. These pages are stored in memory and are relatively easy to examine.
  3. The password is readable on web proxies. Many larger corporations, as well as internet service providers, offer web proxies to allow faster downloads as well as some level of anonymity for web users.

The HTTP specification specifically states that HTTP is not considered to be a secure method of user authentication (unless used in conjunction with some external secure system such as SSL).

It is estimated that between 1 and 2 percent of e-commerce transactions are related to fraud. As customers are becoming more 'net savvy', they are starting to examine web page types and are attempting to only use secure systems. Therefore, any organization that wishes to safeguard its customers' data should start with secure transfers of user login and password information.

Good Practice

A server MUST NOT solicit any passwords in clear text.

Good Practice

A client or browser MUST NOT transmit passwords in clear text.

There are no scenarios where it is possible to transmit passwords in the clear without risk. Every scenario that involves possibly transmitting passwords in the clear can be redesigned for the desired functionality without a cleartext password transmission.

Automatic Protection by User agent

It is tempting to build user agents that refuse to send sensitive data in the clear, or to warn users. There are two problems. Firstly, the user agent cannot determine which data or information is sensitive. Sensitive information is not always input using password masking, often for good reason. Secondly, when the user agent is running arbitrary code, such as when javascript is enabled, a program or script can be used to process a clear text form field (possibly a password) for transfer in many ways, which are too difficult for the browser to analyze.

2.1 Secure transfers

While it's not the purpose of this paper to do an exhaustive description of secure transfer methods on the Web, there are a few common methods used today which are easy to implement;

2.1.1 Digest Access Authentication [Digest]

Digest Authentication acts as an extension to HTTP 1.0 and provides a way for authentication between parties without transmitting the password over the network. Instead the password is treated as a secret input to a digest algorithm. The resulting digest is transmitted and verified by the server.

However, the digest method is often not practicable and has known security weaknesses. The Digest method requires that both parties have access to the same initial secret value. Many systems store passwords as a salted hash, and the result is that it is not possible to use such pre-existing passwords for computing the digest. For example, operating systems that store salted and hashed passwords cannot reuse those passwords for Digest Authentication. The Digest method is subject to dictionary attacks because a single session can be recorded and attacked offline. The Digest method is particularly vulnerable in circumstances where passwords are known to be of insufficient length and complexity to thwart such attacks. Short passwords or those using common words should not be used with digest authentication. Indeed, the sophistication and power of dictionary-based attacks continues to increase such that longer and more complex passwords are increasingly vulnerable to attack. Great care must therefore be taken using digest authentication, and it should be noted that few systems on the Web today require sufficiently strong passwords. The Digest method is also subject to man in the middle attacks because an intermediary can degrade the quality of service to basic authentication.

Given these weaknesses in Digest and password selection, users may erroneously believe the transmitted passwords are secure. Digest should only be used when the costs of more secure systems such as SSL/TLS do not justify the benefits and when strong passwords are encouraged or guaranteed.

2.1.2 Secure Socket Layer (SSL/TLS)

SSL/TLS is a protocol developed for transmitting private channels via the Internet. SSL/TLS works by using a private key to encrypt data that is transferred over the SSL/TLS connection. Most browsers support SSL/TLS and most sites which require sensitive information such as credit card information use SSL/TLS today.

It is important to correctly implement SSL/TLS. Any page soliciting sensitive information such as passwords must be transmitted using SSL/TLS to prevent an attacker from submitting an imitation page that does not use SSL/TLS.

2.2 SOAP Based transmissions

SOAP messages are often sent using HTTP and any SOAP message is subject to similar password security concerns. While SSL/TLS can be used to secure SOAP-based messages point to point, the issue can be more complex if SOAP intermediaries are used. The previous Good Practice advice of "A client or browser MUST NOT transmit passwords in clear text" applies to SOAP messages and so passwords and sensitive information SHOULD be transmitted in a secure manner and not as clear text. If confidential information is to be sent as part of the SOAP package, publishers SHOULD either use SSL/TLS or XML Encryption for sensitive data elements. Further information on security for SOAP messages can be found in Security Challenges, Threats and Countermeasures Version 1.0 [WSI] or on the OASIS Web Services Security TC home page [WSS].

3 Passwords displayed in Browser

HTML allows authors to create input forms. If a form field is a password, password masking SHOULD take place to protect the user from onlookers seeing what is being entered and stop anyone from later using the 'back' button to discover passwords.

<form name="form1" action="https://www.mydomain.com/myform.cgi" method="POST">
    Enter Password : <input type="password" size="25"/>

Good Practice

User agents SHOULD use password masking when passwords are displayed in an HTML form.

This Good Practice does not contain a MUST because there are a few scenarios where password masking is not required. One example is that the user may request that the password be displayed in the clear in order to check the password as it is being entered. Another example is the previous example of a password intended merely to stop web crawling and which consequently is not particularly sensitive. Such non-sensitive passwords may be displayed without masking in addition to being transmitted in clear text.

A References

[IETF RFC 2119]: http://www.ietf.org/rfc/rfc2119.txt

[W3C]: W3C Security Home

[Digest]: HTTP Authentication: Basic and Digest Access Authentication, RFC 2617, Draft Standard, IETF. Available online as http://www.ietf.org/rfc/rfc2617.txt.

[WSI]: Security Challenges, Threats and Countermeasures Version 1.0, WS-I. Available online as http://www.ws-i.org/Profiles/BasicSecurity/SecurityChallenges-1.0.pdf.

[WSS]: Oasis Web Services Security (WSS) TC.