Microsoft officials informed a core group of developers that Windows NT 5.0 will not
ship until the first quarter of 1998. Are you surprised?
In a briefing under nondisclosure held in February at corporate headquarters, Microsoft
experts told a group of about 50 developers that the initial beta for NT 5.0 will not be ready until
late this summer at the earliest. The final versions of NT Workstation and Server 5.0 are now slated
for release early next year, instead of late 1997 as Microsoft had announced previously.
This briefing marks the first time Microsoft has acknowledged that NT 5.0 will be a 1998
product. One anonymous developer said, "I'll be surprised if they make [the first quarter of
1998]. I think [the second quarter] is a more likely bet."
Apparently, Microsoft is struggling with key decisions on how to package NT 5.0, having not yet
announced a date for a release candidate. Microsoft is not publicly revealing any status of NT 5.0
packaging, but the company does seem to be moving forward with plans for two different versions of
NT 5.0: an enterprise version laden with features, and a lighter-weight version. Rumor has it that
Microsoft's top operating system (OS) managers have been questioning top OS analysts about the two
different versions of NT 5.0.
Microsoft officials say they are engaged in figuring out how to release NT 5.0 and what
technologies will be part of the release. Some people speculate that Microsoft will release its
lightweight version first because of unexpected delays in the enterprise version. Insiders say the
enterprise version will include clustering, SNA Server, and the new Transaction Server.
According to analysts, Microsoft needs to break the stigma that NT is a NetWare or OS/2
replacement. One way to accomplish that positioning is to segment the market.
Another advantage of segmentation is that it will let Microsoft focus its marketing efforts.
Analysts also say that many people using NT don't need some of its high-end technologies and that a
departmental-server version of NT would be highly useful and do well as an early release of NT 5.0
Industry experts guess that Microsoft OS engineers are probably still discussing what to bundle
with NT 5.0. Another topic of speculation is that they may not yet have all the technologies
available to put the high-end product together.
—Mark Joseph Edwards
OS/2 Gets an NT Spin
In a move that screams, "If you can't beat 'em, join 'em," IBM recently
announced three components that help OS/2 Warp Server work better with Windows 95 and Windows NT 4.0
clients. The new components include improved administration and security. IBM officials believe
these components give NT Server some OS/2 capabilities, such as integrated backup and recovery and
better scalability, while adding a bit of NT-ness to OS/2.
The new components support multiple user profiles on one system, and systems policies. Two of
the components create a designated home directory for Windows in the OS/2 Warp Server domain where
administrators can set security policies. The third component will for the first time let both Win95
and NT recognize OS/2 as a resource of the Network Neighborhood desktop object. To read more about
IBM in the news, visit http://www.software.ibm.com/news.
—Mark Joseph Edwards
NT and UNIX on IBM Mainframes
IBM's mainframe and midrange operating systems are experiencing an ever-diminishing lead
over high-end UNIX and Windows NT computing. Analysts consider NT as the main threat to IBM, yet
research still shows that UNIX holds a lead over NT for solid enterprise computing. Analysts say
this situation is because IBM's OS/390 supports native UNIX APIs and thus UNIX applications.
IBM's OS/400 won't support NT natively, but the AS/400 has an Intel-based board that will soon
run NT applications in the AS/400 system. The Gartner Group says the OS/390 will support NT
application interfaces this spring and expects IBM to shift to NT as its most strategic,
general-purpose enterprise platform by 2001. For more information, visit http://www.ibm.com and
http://www.gartner.com.
—Mark Joseph Edwards
Tricord Exits Enterprise Server Market
Tricord Systems has announced a significant shift in its corporate strategy. The company
is exiting the enterprise server market to focus on developing and marketing distributed data-access
and data-management products. The company's technology facilitates transparent sharing of
network-attached storage among clustered servers running Windows NT. The technology lets IS
organizations centralize data storage and reduce the costs associated with managing storage
and sharing data.
Since Tricord introduced the concept of the super-server five years ago, the market has changed
dramatically. Technical differentiation, which is what has set Tricord apart from the pack, has
become increasingly difficult as standard server components have become more sophisticated. Intel's
Pentium Pro architecture, which mandates specific bus architecture, is a prime example. This trend
lets a few large vendors control the market with a high-volume, low-cost strategy, so Tricord cannot
compete and stay profitable.
Is Tricord's shift in strategy an abandonment of its core competency? Tricord officials
responded on their Web site by saying, "No. Tricord's distributed data access and management
technology combines the company's recognized expertise in I/O software technology and distributed
file system software acquired from Reliable Distributed Information (RDI) in 1996. This technology
is not a start-up project. It was an integral part of Tricord's Pentium Pro server development and
is well underway. The combination of existing and acquired technology presents Tricord with the
opportunity to be the first company in the Intel/Windows NT Server market to create file-level I/O,
managed at the intelligent controller level instead of in the CPU/main memory environment. The
benefit of this technology is providing the market with Windows NT-attached storage that is easier
to use, has higher performance, and provides system-level reliability across a cluster."
Addressing what this decision will mean for Tricord's future, a company representative said, "Although
there are no guarantees, Tricord is confident that its strategic shift provides greater opportunity
for success than continuing to compete in a server market dominated by commodity products."
This confidence is supported by five key factors:
- I/O technology is Tricord's recognized core competency, and specialization in this area is a
natural focus for Tricord.
- Next-generation I/O technology is not a start-up project; the technology was an integral part
of Tricord's next-generation server program and is well under way.
- The market for NT-attached storage is growing at an annual rate of 31 percent.
- Tricord's software I/O technology implementation offers the opportunity to partner with
numerous key players in the market, including disk controller manufacturers, server companies, and
storage vendors.
- The strategy shift aligns Tricord's current financial strength with an appropriate level of
investment.
- The super-server market presents Tricord with the opportunity to capitalize on its core
competency of I/O technology and introduce a product set beyond just Tricord servers. Tricord will
continue to support its existing user base but will stop manufacturing, designing, and marketing its
hardware at the end of this year. The company's intent is to focus on the I/O performance market
(controller cards, disk subsystems, NT drivers, etc.).
Bob Robinson, CEO of LANology of Houston, Texas, said, "Tricord's exit from this market
leaves NetFRAME in a great position to take up the slack in the microcomputer super-server market.
This move should also create a situation where Compaq and other similar companies can get in and win
former super-server customers. This development may cause many former super-server buyers to
consider purchasing high-end multiprocessor Pentiums and Alphas with faster PCI bus architectures,
or perhaps buy Sequent Servers instead." Read the complete details of Tricord's decision on the
Web: http://www.tricord.com.
—Mark Joseph Edwards
Previous
Register
