Novell Up Close -- What can the once great Novell do
to regain its footing?
By Christy Hudgins-Bonafield
Pray. That's what many top analysts suggest. But after six months researching
Novell and taking the pulse of its chief decision-makers, we're convinced
Novell may yet have a fighting chance not only to retain much of its installed
base but also to take a leadership role in technology innovation.
That chance is premised on the vision and pragmatism of a single man:
Novell's chairman and CEO, Eric Schmidt. Novell isn't headed in the wrong
direction, Schmidt asserts, it just hasn't been moving quickly enough to that
destination. And Schmidt says he is convinced he can remedy that problem
by providing focus and bumping up the pace.
If ever Novell needed a savvy technologist at the balustrades, it is now.
Novell is under siege on all fronts. Historical change, in the form of the
Internet, caught the company unprepared. From one side, it is being
pummeled by the marketing muscle of Microsoft Corp.; from the other, by
standards-based innovation from Netscape Communications Corp. From
within, Novell must counter the complacency engendered by years of
networking leadership. It must shape up, too, after a period of corporate
corpulence and an expansion spree that left it bulging with fatty Unix
technology and superfluous application acquisitions (see "Off Course," on
Network Computing Online at
www.NetworkComputing.com/818/818f2.html)
Novell's chance for success lies in its finding a way to bar Microsoft's inroads
into its base, especially low-end networks, while it goes about the task of
reinventing itself. And Schmidt, Sun Microsystems' former chief technology
officer, is the man in the lab coat (see our interview with Schmidt on page
90). In his vision, Novell continues to derive its primary revenue from the
NetWare platform, with customers drawn to that platform by
revenue-producing directory-enabled services: single sign-on, inventory
management, authentication and commerce. To succeed, Novell must pick up
the pace, lead the competition in directory services and the Internet, and build
the better mousetrap.
One of Schmidt's first actions upon taking the helm at Novell last April was to
almost double the 165 engineers in the Internet Infrastructure group
responsible for Novell Directory Services (NDS) and applications.
Schmidt gives himself two to three years and any or all of $1 billion in cash to
turn Novell around. The timeline is founded on the "network effect"-a belief
that customers will stand by Novell for a certain length of time based on
factors such as loyalty, the risk involved in change, and training and
investment in existing applications.
Significant change isn't apt to surface until the first quarter of 1998, although a
major announcement from the company is expected sometime in November.
Meanwhile, Novell is hard at work tightening customer hooks with low-cost
or free products such as Application Launcher, which help to reinforce its
Internet and directory service directions.
Bandaging the Biggest Wound It is on the low-end that Novell is most
vulnerable. There, the 66 percent of customers still using NetWare 3.x have
been jeopardized by the company's failure to create a migration path for 3.x
servers alongside newer 4.x products. Faced with a forklift migration, with or
without Novell, these users have the greatest incentive to move to the
administration and applications advantages of NT.
Schmidt is close-mouthed about how Novell will hang onto these users, but
he insists that ceding this base to Microsoft and Windows NT isn't an option.
Instead, Novell is apparently working on a range of 3.x migration strategies
and tools. For example, Novell recently announced free distribution of a
migration tool from Simware, and there are strong hints that a lightweight
version of NDS will emerge for the 3.x base. And Novell officials are so
confident in their approach that they predict a 50-50 IntranetWare-3.x split
by mid-1998.
Frank Dzubeck, president of consultancy Communications Network
Architects, says a simpler directory would also shore up Novell's distribution
channel, which has been cut just about in half since its heyday. (Novell
contends this reflects industry consolidation, not Microsoft inroads and that its
channel relationships remain quite strong.)
But analysts agree that no matter how valuable the channel is to Novell, it
can't be the sole foundation for a wholly reinvented company. Schmidt is
aware of this. As a result of his working with sales and marketing staffs during
his first several months at Novell, the channel now will handle direct sales and
support-but Novell will start accompanying its channel partners at user sites.
In other words, Novell plans to take up the Microsoft Gatesian
mode-catering to and evangelizing larger customers. Novell is also attempting
to better categorize the 5,000 some leads it receives monthly to pass along to
the channel through a new Live Leads Web site.
Taking a Cue from Microsoft Sales and marketing isn't the only arena where
Schmidt is tracking rival Microsoft. He already plans to support Microsoft's
ADSI on NDS because he assumes developers will be writing to that
interface. In addition, Novell will support the Java Naming and Directory
Interface (JNDI) it is developing with JavaSoft, Netscape, IBM Corp. and
Hewlett-Packard Co.; and Schmidt is urging that a 12-month delivery
schedule given by JavaSoft can be hastened with incremental API outlays. By
press time, Novell expected to provide at least limited access to JNDI,
acknowledging that the bigger issue will be determining how long it will be
before JavaSoft incorporates JNDI in its Java Development Kit.
Although support for ADSI highlights Schmidt's pragmatism and Novell's
long-held weakness in attracting applications developers, NDS' interfaces
also play upon Schmidt's efforts to "Javatize" Novell. Eventually, he expects
to rewrite the bulk of existing Novell services in Java, with just about
everything-except, perhaps, IntranetWare's kernel-in Java. And if Java is to
emerge in a big way, it will need the kind of security, management and
messaging tools a company like Novell can build.
Paul Zagaeski, senior industry analyst for the Giga Information Group, says
that in his first few months, Schmidt has taken more actions than any CEO in
Zagaeski's experience, certainly more than the string of deliberative top
executives who preceded Schmidt, including Bob Frankenberg and Joe
Marengi.
Schmidt's initial steps included laying off about 1,000 employees and
overseeing the exit of about one-third of Novell's vice presidents. Some
insiders say the message is clear-Novell is a new company and those who
wanted to do things the old way no longer hold sway. Others say that without
additional cuts targeted at high levels, Novell will continue to be a debating
club, rather than an action-oriented company.
Nevertheless, Schmidt obviously commands respect. Some of Novell's top
technologists, like chief scientist Drew Major, say they're pumped up again
and working until the wee hours on projects like BorderManager, which they
believe in and know their CEO embraces. Michael Simpson, director of
marketing for Novell's Internet Infrastructure division, who also spearheads
its NDS drive, says Schmidt is bringing what works for NDS to the rest of
the company, making services cross-platform, focusing on standards and
protocol independence, creating channels and exploring new Internet service
provider (ISP) turf.
But outside Novell's doors, a crisis of confidence clearly exists. Zagaeski says
some analysts at Giga have "written off Novell as an enterprise supplier," an
assessment that, of course, is subject to change. David Passmore, president
of Decisys, also predicts that Novell could shrink to one-third its size over the
next three years-primarily because NT is overtaking it on the NOS side, Sun
and Netscape are way out in front on the Internet, and plenty of competition
is emerging for Novell's end goal of directory-enabled applications. Users
and analysts, too, increasingly are comparing the Novell of today with the
Banyan Systems of two years ago.
Flaw in the Plan? The biggest potential flaw in Novell's plans is its promotion
of NDS as the "other directory" for the industry. Novell's goal is to cede
Microsoft's upcoming Active Directory to pure Microsoft users, while
positioning NDS as the "800-pound gorilla," that is, the enterprise directory
for companies with Microsoft and non-Microsoft operating systems.
Novell's strategists "just don't get it," says a strategic planner for a large
telephone company's internal network. Directories haven't even hit their stride
yet as an integral part of business infrastructures, but Novell is assuming that
they will become so critical to the next generation of technology that their
bundled, or commodity nature, is guaranteed.
Even though the company enjoys a technological edge with NDS, most
Novell watchers believe it decided too late to move NDS onto other
operating system platforms and to offer NDS interfaces as Internet standards.
The prevailing sentiment is that users will prefer to wait for Microsoft's Active
Directory on 5.0, even if it takes a year or more for it to hit the streets. And
what they'll get is a directory that is native to NT, essential to Cairo and
accessible over the Internet via the Lightweight Directory Access Protocol
(LDAP). On the other hand, even Novell's closest NDS partners don't seem
to see NDS as the directory to rally around for their own products. They're
offering NDS as an option alongside their own directory strategies-with
Active Directory as yet another possible option in the future. Because of this,
many Novell observers-though not all-say it's clearly in Novell's interest to be
OS-independent as well as NDS-agnostic.
Once Active Directory is available, Forrester Research's Jon Oltsik says, 70
percent of the server base will simply shift to it. Novell won't win if it tries to
be the 800-pound directory gorilla, Oltsik says. "The game isn't to
outdirectory Microsoft, it's to figure out how to work with Microsoft," he
explains. "Users are looking for standards to make directories interoperate
and NDS does nothing there."
Schmidt, however, is convinced that what is occurring is early market
positioning and that once the dust settles-and the rest of the industry truly
appreciates how difficult it is to develop a good directory-NDS will win.
After all, NDS development began more than nine years ago, and it's taken
almost seven of those years to achieve a product Novell deems truly stable.
But what if Schmidt is wrong? If the world remains multidirectory, Novell's
NDS emphasis will detract from efforts it could be pursuing if it had a clear
vision of its mission to provide applications and infrastructure for overarching
metadirectories. But several factors mitigate the damage if Novell is wrong.
For one, Novell's plans call for supporting a variety of directories within
NDS-if not at a metadirectory level, then at least on a synchronization basis.
For another, APIs from metadirectory providers like Zoomit can be used at
some future date to create metadirectory functionality atop existing directories
like NDS. Finally, if JNDI is widely accepted, upcoming directory
applications written to it by Novell should eventually run on any JNDI
directory.
Of course, Novell's NDS emphasis also leaves it open to other major
networking players beating it to the metadirectory punch. Microsoft has said
that it doesn't plan to be in the metadirectory business, but Netscape officials
are publicly embracing the concept-leading to the distinct likelihood of a
Netscape-Zoomit marriage. Such a move would recognize the fact that it will
be years-if ever-before LDAP and its directory APIs have sufficient
functionality and popularity to replace today's many directories and access
protocols.
Edwards Reed, Novell's chief architect for directory and security services,
says Novell will certainly aid metadirectory providers, like Zoomit,
WorldTalk and even Netscape (if it goes into that business), but that Novell
shouldn't provide its own metadirectory. Selling shrink-wrapped
metadirectories becomes a highly consulting-oriented business, Reed says. It
would leave Novell "chasing" the directory release cycles of every other
company, when it needs to focus on its strength-building the mission-critical
directory that, in turn, can be integrated with a metadirectory. Simpson adds
that Novell will synchronize with NT domains (this capability is already
shipping), Microsoft Exchange, Notes and perhaps one or two other key
applications. It also will provide a software developers kit for its event system
so that third parties can build meta synchronization products.
The Tactical Plan Before Novell can execute its master plan, however, it must
be able to perform two key tasks: keep existing customers on board and
satisfied, and migrate as many of those customers as possible to a base with
directory services.
One obvious Novell tactic, then, is to make it easy for Novell users to extend
their Internet technology reach. BorderManager, which shipped on time in
August, is the first salvo in this effort. Novell hopes BorderManager will
define a new directory-based type of Internet/intranet middleware. Instead of
positioning infrastructure security, manageability, filtering and performance on
the server or the browser, Novell wants to package these capabilities in the
middle. The idea is to off-load the client and server while creating a single
point of support for products that must otherwise be integrated by the
customer.
Novell's Major says this middle infrastructure will be what takes the Internet
to the next level. One of its chief advantages is central administration-for
example, implementing security policies in the middle, instead of at the
browser, where users can disable those efforts.
If he could push a button that would escalate the pace of anything at Novell,
Major says it would be BorderManager-"which smells to me like resource
sharing smelled 15 years ago. Everyone needs it, but they just don't know
that they need it yet. I see the same thing happening with this that I saw
happen with NetWare. We had a vision and added new functionality on it for
10 years."
Initially, BorderManager will address proxy caching, firewalls, gateways from
IP-to-IP and IPX-to-IP, and VPNs. Simpson compares the application
package to Notes and sees it selling IntranetWare the same way Notes sold
OS/2 (for an interview with Novell's Simpson, see Network Computing
Online at www.NetworkComputing.
com/818f2/f2.html). Major adds that "a lot of BorderManager will show up
cross-platform," even though Novell is still weighing factors such as the
slower speed of services on those platforms.
Novell also apparently is studying what role it might play as a server provider
for network computers based on Java, though Schmidt isn't yet willing to
discuss plans. Major says he believes it would be a good business for Novell.
Native support for IP, alongside optional IPX, is yet another Internet
migration strategy emphasized by Novell with project Moab, the new more
fault-tolerant IntranetWare slated for late 1997 or early 1998 delivery.
Coleman Barney, vice president of Novell's Internet Access division, says
Moab will include some "interesting twists to the stack." Clustering technology
is another plus in Novell's efforts to maintain its base, with most analysts
seeing the defection of key Wolf Mountain engineers as more of a political
problem than a true setback.
How Wise GroupWise? Whereas some Novell watchers, like Forrester's
Oltsik, are convinced the company should sell its collaborative GroupWise
technology, most analysts and Novell officials disagree. GroupWise-with its 7
million users and 75 percent compound annual growth rate-is a moneymaker.
It's also an obvious fit with the directory enablement strategy of Novell
(GroupWise 5 is based on NDS and runs on NetWare 3.x, 4.x, NT and
Unix, though a NetWare server is needed for NDS). Oltsik argues that it's a
losing battle to fight "Lotus, Microsoft and the Web," but Novell officials hint
that GroupWise-NDS may find its way into an ISP package. Major says
Novell's goal is to provide applications choices-like GroupWise, Lotus Notes
and Netscape's collaborative services through the partnership of Novell and
Netscape in Novonyx (see "Novell and Netscape Give Birth to a New
Venture," June 1, page 24). Applications will have to stand on their own, he
says, and so far, GroupWise is doing just that.
Novell touts network management and administration as another strong facet
of its directory and Internet strategy. Novell's project Houston, for example,
is a pure Java environment that rolls together application tools such as
Novell's ManageWise, NetWare Administration and third-party applications
to run on any OS and be accessible from any Java device. The environment
will rely extensively on NDS coupled with roles-based authentication to
authorize and segment management tasks across a network. A software
development kit and beta product are expected by year's end with delivery in
the first quarter of 1998.
According to Reed, Houston is "the Java framework that lets us manage the
content of the directory and the directory topology." Houston also will be
critical in determining where replication occurs and where replicas are stored.
Reed says the protocols are now being documented as LDAP protocols, so
the NDS directory can be managed through LDAP.
Todd Chipman, a product line manager for ManageWise, says Novell also is
evaluating a 16-bit application version of ManageWise 2.1 on NT. In
addition, a ManageWise IP product will be ready soon-perhaps by
mid-1998. Yet another management tie-in to NDS is Novell's free
Application Launcher, which relies extensively on NDS for the configuration
information used in software distribution. Novell's managers also are trying to
drive an NDS wedge into enterprise management systems from Computer
Associates International and IBM, but talks with IBM remain preliminary and
CA seems to be resisting anything much more than promoting NDS for
Novell environments.
The NDS Plan In recent months, Novell advanced the NDS Common
Services Engine for companies shipping NDS on their own independent
OSes. The engine introduces modularity so that services can be added to
NDS-including authentication or advanced print capabilities-without requiring
a new shipment of the operating system.
The engine is expected in the first shipment of NDS-supporting products from
Sun and Novell for NT, and in the second shipment of products from IBM,
Fujitsu, HP and The Santa Cruz Operation. Novell estimates that these
second-round shipments will begin in about six months. NDS is provided
without charge, but OEMs are charged a fee for replication-probably under
$1,000 per server. The OEMs may then derive their own income from the
service.
Simpson says Novell is also pushing to see NDS used in applications, through
deals like those involving Oracle Corp.'s Network Computing Architecture,
which lets Oracle developers use NDS transparently. In-house, too, he says,
Novell is remedying key problems like the fact that no one from developers
services was devoted to NDS last year.
Novell, for example, recently launched a Java-NDS developer training/prize
promotion featuring a custom Harley-Davidson motorcycle, pizza and Jolt
Cola supplies.
Simpson also expects JNDI to be a significant draw for application
developers. While JavaSoft officials have publicly said the interface may take
up to a year to complete, Simpson says JNDI should achieve stability
sometime this fall, with "a bunch of applications" emerging by early next year.
Several Novell officials also say Novell will be LDAP's biggest cheerleader,
deploying the richest implementations of the evolving standard protocol-but
also recognizing, Major says, that NDS on the back end will provide better
fault tolerance, security and manageability than a traditional LDAP server.
Novell is also pulling applications, like Application Launcher, into Simpson's
division, and apparently is looking closely at Entrust for the certificate
authority services that play off its directory.
The Leader Of course, the big question is: Can Eric Schmidt pull off the
tactical and strategic changes that need to occur at Novell? Many insiders
believe he can. They're amazed at Schmidt's ability to answer 90 percent of
shareholder questions personally and at his intuition about people. Aggressive
people within Novell are now the rule, says Simpson, instead of the exception
to the rule.
"The way Eric feels is, 'Screw playing it safe. Make mistakes. Just do
something.' You can fix mistakes, but you can't fix the fact you aren't doing
anything. Novell's biggest problem was not making decisions," Simpson says.
Will Novell ever command 68 percent of the market again? Even Major says
what the company needs to do is grow the overall market, not shoot for a
specific share. And if anyone can bring Novell out of the doldrums, it's Eric
Schmidt, he says.
"Novell is fun again," Major said after working until 3 a.m. the previous night.
"We've started building new things, instead of fixing things."
======================================================================
I Can't wait 2 to 3 more years DJ. I hope they can do it.
|
