In a lingering saga that may one day lead to the standardization of an open email encryption standard, PGP has moved to the fore and left RSA Data Security to play catch-up with its currently de facto, yet proprietary standard.
Earlier this year RSA Data Security submitted the S/MIME (Secure Multipurpose Internet Mail Extension) specification to the Internet Engineering Task Force's initial standards review process. However, since then the company has refused to make important concessions that would further its prospects for becoming an IETF-approved standard, said Jeff Schiller, the IETF's security director.
"The problem with S/MIME is that it was developed outside the IETF, and RSA needed to turn it over. They never agreed to do that," said Schiller. "When PGP's marketing people came out and said they were going to have the IETF standard, that woke up the RSA people. But when they thought they could get away with ignoring us, that's what they did."
Currently, the IETF appears to be ready to accept a no-strings-attached proposal by Pretty Good Privacy that may lead to the creation of OpenPGP, an alternative, but similar technology to S/MIME.
If accepted by the IETF, the OpenPGP proposal could create a situation where two competing standards will both become popular on the Internet, despite the fact that interoperability may be difficult to achieve. If OpenPGP moves into a working group, it could become an official standard by the end of 1998, said Schiller.
RSA maintains that it is still interested in creating an IETF-blessed standard, and is willing to submit a charter if it's able to retain intellectual property rights, which is allowed in some circumstances under IETF guidelines.
"There were some requirements made on S/MIME and there's been some misunderstanding on what was expected. But there's nothing about S/MIME that should preclude it from moving ahead as a standard," maintains Steve Dusse, RSA's chief technology officer.
Schiller said RSA wanted to provide the "building blocks" of the S/MIME specification, but retain intellectual property rights to certain parts of it, which would allow RSA to retain licensing rights to the standard should it be approved. Meanwhile, the PGP application appears to have no such proprietary provisions, and the IETF prefers to use "unencumbered" technologies if they are available.
Although OpenPGP and S/MIME have similar ends, the means in which they implement public-key-based encryption are different. S/MIME relies on a stricter, directory-based method of issuing, distributing, and managing the keys that identify users. This often entails going through a third party like VeriSign, which RSA has business partnerships with. The PGP model relies more on a trust system, whereby users pass keys among themselves, and the onus of matching someone's identity with their key is on the individual users. Other differences involve the complexity of the security algorithms used, and some sources said the IETF is more inclined to go with PGP because they implement stronger algorithms.
"PGP specs require implementation at a minimum 128-bit security, and that hinders deployment for export," said Eric Berman, product manager for Microsoft's Outlook Express, referring to the US government's restrictions on the export of strong encryption products.
But just as Netscape and other vendors have been able to capitalize on proprietary technologies that turn into de facto standards, RSA may have enough of a lead in the corporate market to entrench itself well before the OpenPGP standard is approved. Microsoft and Netscape already support S/MIME in their email and browser products, and may or may not choose to support OpenPGP if it's approved.
"Our corporate customers are deploying x.509 infrastructures [used in S/MIME] and are basing secure email with it. Most aren't going in the direction of PGP," said Jim Reitz, a server product manager at Microsoft. "Being a standard is one thing, getting customers to demand it is another," he added.
The end result may be that two different email encryption technologies are standardized, leaving users to sort out which crypto suits them best.
