![]() |
|
| C O L U M N | |
The Strategy Behind the Directory May 15, 2000 By: Ron Anderson Once upon a time, there was a large company that was like a lot of other companies around the world. It had a centralized IT organization running many of its business-critical functions and a distributed IT infrastructure designed to meet the special needs of its various parts. In 1994, the company was running Novell NetWare 3.12 for file and print services and had been experimenting with NetWare 4 for a year. The NetWare administrators in central IT knew that the company would benefit from the consolidated management infrastructure provided by the new directory services in NetWare 4. Therefore, when 4.1 shipped they began to move their production 3.12 servers along with a couple of thousand users to NDS. It took a few months for the NetWare administrators to move past their flat-file bindery mentality to realize the efficiencies afforded by the new hierarchical infrastructure. For the next couple of years they preached the gospel of directory services to the comparatively large group of distributed NetWare administrators in the company with a great deal of success. By 1997, they had 97 percent of the NetWare users in the company in a single tree and had setup a distributed administration model that didn't threaten anyone's turf. The tree had about 6,000 user objects and represented nearly all the employees in the company. But that's as far as it went. Over the years, NDS proved itself to be scalable and reliable. However, though NDS let the company manage NetWare resources and users more efficiently than it could with little bindery islands, NDS didn't act as a strategic resource and didn't help the company take its business to the next level. The NetWare administrators had a vision for a unified directory of users, resources (including IP addressing and DNS entries), Internet services (such as caching and firewall protection) and policies with all of the business's critical processes tied-in, but the vision never materialized. The failure to continue to move forward with directory services for this company had very little to do with the technology itself. Novell introduced a multitude of directory-enabled services--single sign-on, BorderManager, DNS, DHCP, Internet Caching System, ZENworks, NetWare Enterprise Print Services, and Certificate Server. NDS continued to mature and demonstrate even greater scalability and speed while divorcing itself from any reliance on NetWare with native cross-platform implementations and standards-based access methods. However, the more the technology changed, the more the infrastructure stayed the same and the company could not take further advantage of NDS' enhanced functionality. So how can you avoid the trap of failing to move your directory into the strategic position it needs to occupy within your company? First, approach directory services as a solution to business problems, not as a solution to technical issues. Because you understand its technology you'll be the first to glimpse the promise of directory services. You'll need to relate that promise to the bottom line to get the top-down support you'll need to make directory services strategic. Remember, you don't just need to convince your boss, you need to convince your boss's boss and his boss, clear to the top. Directory services need to be viewed as a strategic component in a healthy enterprise-computing environment. In fact, I believe that a well-implemented directory service is the key to pulling the mishmash of disparate systems and services that exist in most organizations into a unified, manageable whole that can add value to any company's main line of business. And now that Microsoft has entered the directory fray with its Active Directory, companies that missed out on NDS because they were running NT rather than NetWare will be starting to make plans for implementing their own directory service. Microsoft and Novell have sponsored third-party studies to help you get started.
Second, consult with your software vendors and let them know that your directory is strategic and that they'll need to work with you to fit their products into your infrastructure. There aren't many things that are worse for an IT organization than having to support multiple user databases. It's a big drain on IT resources, a huge security problem and a hassle for your users. It also makes it difficult to keep all the pieces in sync. If everyone in the company, from top to bottom, is telling your vendors the same thing, the vendors will get your message. Third, don't implement any new in-house systems that don't leverage your directory. It's always tempting to do things the way you have always done them because doing so is faster, doesn't require your systems people to adopt new methodologies and doesn't require retraining. Prepare yourself for a painful process because your colleagues probably won't be on the same page when you start down this path. Remember that they have a big professional and personal investment in technologies that may or may not fit into a directory-based infrastructure. Once again, the top-down buy-in is critical for pushing this sometimes-painful process through to the point that it becomes policy. The payoff will be a cohesive, manageable infrastructure that adapts well to change, provides for distributed administration and provides reliable, scalable services. Finally, plan to integrate your customers and business partners into your directory. Even if you think it will be a couple of years before you can implement customer and partner integration, plan this phase as you begin to design your directory service. It will make your implementation efforts easier in the long run. After all, you'll probably want to do this sooner than you think anyway. Integrating business partners and giving them an "identity" is going to be a critical feature for B2B and B2C functions. This includes customized services for customers and vendors based on who they are (identity). This class of services requires directory services for managing the customized experience. Doing so will become a requirement in the near term, so don't let it catch you by surprise. By following such business-first practices, our fictitious company and yours will live happily ever after. The end.
Send your comments on this column to Ron Anderson at randerson@nwc.com.
| |












