By Jeff Sweat
Microsoft Corp has abandoned the Object File System (OFS) that was to be the centerpiece of Cairo in favor of an enhanced version of the existing NT File System (NTFS), according to our sister publication ClieNT Server News. Microsoft low-keyed the news and practically lost it amongst the NetWorld+Interop hype over the launch of its Exchange messaging suite. In place of what might have been a radical piece of technology, if it could have pulled it off, the still-distant Cairo will add features to NTFS that Microsoft claims will achieve the same goals as the mythic OFS. Microsoft’s lead NT product marketing manager Mark Wood said that, unknown to the world-at-large, OFS was stripped of its objects-only meaning long ago and the term OFS actually described a feature set rather than a specific product. The objectives for Cairo haven’t changed, he said, only the technology taken to reach them. Since OFS was first announced in very sparing detail, as Microsoft admits, there’s been a lot of brainpower applied to connecting the dots and creating pictures of what it should look like by outsiders, Wood said. With little to go on, they drew pictures different than the ones in Microsoft’s head. Then when Microsoft made more details public, those pictures were at variance with the so-called reality. Microsoft intended to create a whole new object system at first, but started casting around for alternative methods early in development after realizing, it contends, that OFS would have been too revolutionary. Wood claims Microsoft had to choose between a forklift upgrade that would have been extremely costly for users as they switched to a new file system or gradual upgrades with the technology layered on top of NTFS. There was also the cost and time of development to consider, probably more so as Cairo slipped. Wood said Microsoft decided it didn’t need an alternate file system because – with enhancements – NTFS would support OFS features like content indexing and queries, extensible attributes and the creation of a logical view of the network that’s independent of its physical configuration – a big disk approach also known as a Distributed File System (DFS). Wood, mindful of the OFS experience, warned not to put too much stock in the DFS name lest its meaning change again and again just like OFS’. There is of course the camp that was never satisfied with the way Microsoft defined OFS and took it to mean that it either never really existed or that the effort failed or that it never worked and Wood’s explanation doesn’t really negate any of those suppositions. The switch raises the spectre of yet another Cairo delay.
Robbing Cairo
Microsoft has finally given up on the fantasy that we’ll see a Cairo beta this year and now says 1997, with product possible in 1998, just as we’ve always maintained. However, Gartner Group analyst David Smith, a 1997-98 man himself, said the substitution may actually speed Cairo to market. As Cairo is pushed off, Microsoft is rolling out pieces of Cairo functionality sooner, to wit, the Windows 95 GUI and Distributed OLE or DCOM, Universal Data Access (the old OLE DB) and now the directory. The directory service shipping with Exchange Server is now to be understood as Cairo-developed technology that will be integrated into NT 4.0. The Exchange directory, which builds on top of NT Directory Services (NTDS), adds a hierarchical X.500 naming scheme and DNS support. It also has an extensible structure that allows users to add attributes such as mailing addresses, fax numbers and names to files. Provocatively, Smith sees the change in file systems as a sign of Microsoft’s further – possibly passionate – embrace of the Internet, since the new directory will be subordinate to the Internet’s Domain Name System (DNS), the stuff that URLs are made of. Since the Exchange directory is closer to what exists now, it will be easier for NT users to migrate to Cairo than if it were OFS-based. On the other hand, stripping out its features robs Cairo of much of its mystique, a P
R challenge Microsoft will doubtless find a way to surmount.
