Is There Light At The End Of The E-Mail Tunnel?

Many of the original developers of shared-file LAN e-mail applications, including Lotus, Microsoft and Novell, have made the move to true client/server messaging environments. These systems promise better performance, better scalability and enhanced support for Internet standards. For the most part, they deliver on this promise, but installing and maintaining these systems is not only fraught with peril, it also may require a total rethinking of the underlying LAN infrastructure. For example, some organizations have abandoned NetWare in favor of NT as a strategic NOS platform, mainly because Microsoft Exchange is dependent on NT's domain-based account system. Some have even survived to tell the tale. Finally, many sites with long histories of reliance on Internet technologies have opted for standards-orie nted e-mail systems. Until recently, that meant building your e-mail environment around SMTP and POP. The biggest advantage of these technologies lies in their relative simplicity. It's been fairly easy for application developers to incorporate support for SMTP and POP into their products. Unfortunately, while SMTP has been enhanced over the years to support features like MIME for binary attachments, POP continues to suffer from its lack of support for location-independent e-mail access. IMAP to the Rescue? IMAP was developed in 1986 at Stanford University with a feature set designed to overcome the limitations of POP. But it languished in relative obscurity for years. Many of us were excited by its capabilities, but also wary that the number of commercial implementations of IMAP clients could be counted on one hand. Now in its fourth generation, IMAP provides a standards-based client/server messaging environment that offers e-mail functionality competitive with proprietary systems like Notes, Exc hange and GroupWise. No longer simply the subject of discussion among Internet mail gurus, IMAP commands the attention of major industry players. Without a doubt, Netscape's decision to build IMAP support into Communicator and SuiteSpot provided the biggest boost to IMAP's credibility. And even Microsoft, where the corporate motto could read "We're committed to open standards--when all else fails," has built IMAP support into its Outlook and Outlook Express e-mail clients. But don't let the hype over IMAP cloud your judgment. While basic interoperability between different vendors' client and server implementations has been established, migrating an enterprise mail infrastructure to IMAP is still risky. An excellent report on IMAP from The Burton Group concludes that it will be a year and a half or more before IMAP and a number of companion protocols needed to realize its potential are widely available in the market. However, IMAP's appeal is compelling, and there can be little doubt that it will emerge as the dominant messaging standard. While Lotus, Microsoft and Novell are on firm ground in touting the superior features of their proprietary messaging and groupware environments, all have moved with a sense of urgency over the past year to build at least minimal support for IMAP into their offerings. To ignore IMAP would be to risk losing market share to companies like Netscape that are betting the house on standards-based systems. So why exactly is everyone so excited about IMAP? It provides multiple modes of operation--offline, online and disconnected--that meet the e-mail needs of millions of users of commercial ISPs as well as millions of users in organizations where access to one's e-mail from multiple locations is essential. People who have struggled with POP-based mail will embrace IMAP, while users of mature but proprietary e-mail systems will find themselves giving up little, if anything, by using a standards-based product. And IMAP effectively decouples the e-mail client from the e-mail server, an element of interoperability that many organizations will find very appealing. Is Soon Quick Enough? Unfortunately, it's not all here yet. Think of IMAP as a work in progress--a solid foundation is in place but there's still an awful lot of mud surrounding it, and it's very easy to get stuck in the mud these days. Detractors first point to the absence of essential elements, which limits the promise of location-independent access. Although it's possible to access one's message store from multiple systems--and yes, even from different vendors' client applications--the lack of a network-based repository for client application preferences and personal address books is a serious shortcoming. Debate continues concerning the appropriate technology for delivering this capability. Most IMAP purists favor the emerging IETF-standard ACAP (Application Configuration Access Protocol), which is designed specifically to address this need. Another contingent, led by Netscape, makes a strong case for schema exten sions to LDAP-based directories. Both positions have merit, and both may be accepted as viable alternatives, but this debate slows IMAP's adoption process. The second major issue slowing IMAP adoption relates to scalability. From a vertical scalability standpoint, there are concerns that the resource requirements associated with IMAP's server-based mail store will be excessive--certainly considerably more than for POP. Converting a single-server environment to one based on multiple servers can present significant technical challenges to system administrators, including changing user e-mail addresses and implementing an address-aliasing system. Horizontal scaling is also a challenge, and while Carnegie Mellon University's Project Cyrus is addressing many of the major scaling issues, it's very immature technology. Finally, there's the momentum of a massive installed base of Notes, Exchange and GroupWise users. These products provide groupware functionality that IMAP lacks. Many system designers are asking just how valuable that additional functionality is, but for now, IMAP's threat to the big three is minimal. And with Lotus, Microsoft and Novell building IMAP functionality into their servers, managers responsible for those systems are likely to stick with what they've got. But for those of you who haven't bit at one of the proprietary systems, a long-term commitment to IMAP makes a lot of sense. Dave Molta is director of network and system services at Syracuse University. He can be reached at dmolta@nwc.com.

Other Columnists Corporate View By Robert Moskowitz On The Edg e By Art Wittmann