Over the past two years, we have participated in industry discussions through the IETF about how to achieve the goal of interoperability among instant messaging networks. At the same time, we have resisted efforts by our competitors to impose a "quick fix" system that would jeopardize our members' privacy and security. Both of these decisions have been guided by two bedrock commitments: To provide consumers with a secure, private, and convenient online experience; and to help build a medium we can all be proud of.

In response to your call for instant messaging ideas to be submitted by today, we are reaffirming our consistent commitment to interoperability with the release of our proposed architectural design for a worldwide instant messaging system. We think that this submission represents a significant first step toward developing more detailed protocols for implementing the kind of full interoperability that we all would like to see.

The IETF knows better than anyone how complex and difficult an engineering challenge is posed by developing standards that protect the consumer. Ultimately, our first commitment and biggest concern in this process continues to be protecting the privacy and security of consumers. Once protocols are published, they will be used by hackers and spammers as a roadmap to plan their attacks. We believe that it is critically important not to release such proposals until we are certain that the security precautions in them are sufficient to protect consumers. If we do not move deliberately through this process, we risk undermining our efforts to serve consumers well. Nobody wants to see instant messaging interoperability made useless by a barrage of offensive spam, attempts to defraud, and virus proliferation.

The design we have submitted is a major advance toward developing a system that forestalls these threats with a server-to-sever approach to interoperability that would work in a manner similar to Internet e-mail. We believe that this approach protects the user's privacy, security and ease-of-use as well as promoting continued long-term competition and innovation in the industry and providing the greatest degree of scalability. Among the advantages of this system are:

• Full Interoperability: The server-to-server architecture of the AOL design allows users of any two IM networks that use the same protocols to communicate with one another at any time;

• Privacy and Security: Under this architecture, users would need to be registered with only one instant messaging system, and would not be required to share passwords, log-in IDs, or other confidential information with anyone outside the network they choose. In addition, the design includes requirements that IM data can't be easily spoofed or replayed by a third party; that messages can't easily be intercepted or hijacked; and that more advanced security measures such as end-to-end encryption or signing can be layered on top of initial implementations. Finally, it requires that individual networks be allowed to use firewalls or other precautions to ensure the highest possible degree of security;

• Scalability: This design would enable the development of any number of instant messaging systems, from large networks like AOL to individual families with their own servers;

• Independence: No government or other central authority would be required to administer the system; and

• User Name Consistency: Users of instant messaging networks would be able to keep their existing screen names or addresses, even if different users had identical names on different networks.

In addition to openly publishing our design, we are publishing this document as an Internet Draft and submitting it to the IETF for consideration as an informational RFC, in the hope that it will serve as a helpful guide to the standards-development process, and that other participants in the instant messaging working group will offer useful comments on this design as we work to develop protocols for its implementation.